Skip to main content

event and task manager

Project description


This is the etm user manual. Further information about etm is available at github and in the etm discussion group at - note especially the files folder. Brief how-to videos are available at youtube.




etm offers a simple way to manage your events, tasks and other reminders.

Rather than filling out fields in a form to create or edit reminders, a simple text-based format is used. Each reminder in etm begins with a type character followed by a brief summary of the item and then, perhaps, by one or more @key value pairs to specify other attributes of the reminder. Mnemonics are used to make the keys easy to remember, e.g, @s for starting datetime, @l for location, @d for description and so forth.

The 4 types of reminders in etm with their associated type characters:

  • task: -
  • event: *
  • journal: %
  • inbox: !

See Item Types for details about these item types and Options for details about possible attributes.


  • A task (-): pick up milk.

      - pick up milk
  • An event (*): have lunch with Ed [s]tarting next Tuesday at 12pm and [e]xtending for 90 minutes, i.e., lasting from 12pm until 1:30pm.

      * Lunch with Ed @s tue 12p @e 90m
  • A journal entry (%): a favorite Churchill quotation that you heard at 2pm today with the quote itself as the [d]escription.

      % Give me a pig - Churchill @s 2p @d Dogs look up at
        you. Cats look down at you. Give me a pig - they
        look you in the eye and treat you as an equal.
  • A task (-): build a dog house, with component [j]obs.

      - Build dog house @j pick up materials @j cut pieces
        @j assemble @j sand @j paint
  • Inbox (!): meet Alex for coffee Friday. This can be changed to an event when the time is confirmed by replacing the ! with an * and adding the time to @s.

      ! Coffee with Alex @s fri @e 1h

    This inbox entry will appear on the current day in agenda view until you make the changes.

  • An appointment (event) for a dental exam and cleaning at 2pm on Feb 5 and then again [@+] at 9am on Sep 3.

      * dental exam and cleaning @s 2p feb 5 2019 @e 45m
        @+ 9am Sep 3 2019

text entry versus forms

  • Text entry removes the need to hunt for and click in the relevant entry box and allows you to keep your fingers on the keyboard.

  • Text entry supports the full flexibility of the superb Python dateutil package. Consider, for example, creating a reminder for Presidential election day which repeats every 4 years on the first Tuesday after a Monday in November (a monthday falling between 2 and 8). In etm, this event would be

      * Presidential election day @s nov 1 2020 @r y &i 4 &M 11
        &m 2, 3, 4, 5, 6, 7, 8 &w tu

    Try this with a form based calendar application.

unobtrusive and timely entry assistance

When you want to create a new reminder or edit an exiting one, etm opens an area at the bottom of the screen that is divided into two parts by a horizontal line. The lower part is the entry area where what you type appears. The upper part is the prompt/feedback area where etm responds to your typing. This response might take the form of providing a suggestion about alternatives, information about the type of input required or feedback about how your current entry is being interpreted. It is important to realize that none of this interferes with your typing - you can blaze away as quickly as you like or even paste complete entries and never glance at the prompt if you like. This is the unobtrusive part of the prompt/feedback process.


just in time entry prompts and feedback

Let's create the election day reminder to illustrate the timely part of the process. Begin by pressing N to create a new reminder and notice that etm automatically prompts you for the item type character and suggests the alternatives.

    item type
    Choose a character from * (event), - (task), % (journal)
    or ! (inbox)


Notice that you don't have to search for the part of the form in which you specify the type of reminder - etm effectively makes the one entry area the appropriate spot for this and all other input. Now enter an * to create an event and etm will prompt you for the event summary.

    brief item description. Append an '@' to add an option.


Enter the summary followed by an @ and etm will automatically display the required and available @-keys.

    required: @s (start)
    available: @+ (include), @- (exclude), @a (alerts),
    @b (beginby), @c (calendar), @d (description),
    @e (extent), @g (goto), @i (index), @l (location),
    @m (mask), @n (attendee), @o (overdue),
    @r (repetition frequency), @s (start), @t (tag),
    @u (used time), @x (expansion), @z (timezone)
    * Presidential election day @_

You see that @s is required, so add the s and etm will prompt you for the value for the start option and describe the type of input required.

    starting date or datetime

    * Presidential election day @s_

As you enter the datetime, etm will display its interpretation of your entry.

    Sun Nov 1 2020

    * Presidential election day @s nov 1 20_

Now append an @ to see the prompt for options and notice that @s, having been provided, is no longer listed.

    available: @+ (include), @- (exclude), @a (alerts),
    @b (beginby), @c (calendar), @d (description),
    @e (extent), @g (goto), @i (index), @l (location),
    @m (mask), @n (attendee), @o (overdue),
    @r (repetition frequency), @t (tag), @u (used time),
    @x (expansion), @z (timezone)
    * Presidential election day @s nov 1 20 @_

We want to repeat so enter an r and etm will describe the required input.

    repetition frequency
    repetition frequency: character from (y)early,
    (m)onthly, (w)eekly, (d)aily, (h)ourly or

    * Presidential election day @s nov 1 20 @r_

Add a y for yearly. The repetition attribute, @r, has several possible attributes which apply specifically to it. Such attributes use & rather than @ as a tag. Enter an & to see the prompts for repetition options.

    repetition &-key
    repetition &-keys: &i (interval), &m (monthdays),
    &E (easterdays), &h (hours), &M (months),
    &n (minutes), &w (weekdays), &W (week numbers),
    &c (count), &u (until), &s (set positions)

    * Presidential election day @s nov 1 20 @r y &_

Enter an "i" to see what "interval" means.

    Interval requires a positive integer (default 1) that
    sets the frequency interval. E.g., with frequency w
    (weekly), interval 3 would repeat every three weeks.

    * Presidential election day @s nov 1 20 @r y &i_

We want to repeat every 4 years so enter 4 followed by another & to see other options. [The display for repetition &-keys is unchanged from above and will be omitted hereafter.] Now enter an M to see what "months" means.

    months: a comma separated list of integer month numbers
    from 1, 2, ..., 12

    * Presidential election day @s nov 1 20 @r y &i 4 &M_

We want November, so enter 11 followed by an & for other options. Then enter an m to see what "monthdays" means

    monthdays: a comma separated list of integer month days
    from  (1, 2, ..., 31. Prepend a minus sign to count
    backwards from the end of the month. E.g., use  -1 for
    the last day of the month.

    * Presidential election day @s nov 1 20 @r y &i 4 &M 11 &m_

We want a monthday between 2 and 8 so enter these monthdays followed by "&" for options. Then enter w to see what "weekdays" means

    weekdays: a comma separated list of English weekday
    abbreviations from SU, MO, TU, WE, TH, FR, SA. Prepend
    an integer to specify a particular weekday in the month.
    E.g., 3WE for the 3rd Wednesday or -1FR, for the last
    Friday in the month.

    * Presidential election day @s nov 1 20 @r y &i 4 &M 11
    &m 2, 3, 4, 5, 6, 7, 8 &w_

We want a Tuesday, so add tu to complete the entry. To check the entry, press ^R to show repetitions

        ┌─────| First 5 repetitions |──────┐
        │                                  │
        │ from Sun Nov 1 2020:             │
        │   Tue Nov 3 2020                 │
        │   Tue Nov 5 2024                 │
        │   Tue Nov 7 2028                 │
        │   Tue Nov 2 2032                 │
        │   Tue Nov 4 2036                 │
        │                                  │
        │           <    OK    >           │
        │                                  │

These are the first 5 repetitions on or after Nov 1 2020. Notice that the start value of Nov 1 2020 is not one of the repetitions since it doesn't satisfy the requirements. These appear correct, so press ^S to save the reminder.

General observations:

  • The general structure of this reminder is

      	* Presidential election day
      		@s nov 1 20
      		@r y
      			&i 4
      			&M 11
      			&m 2, 3, 4, 5, 6, 7, 8
      			&w tu

    The @s and @r entries provide attributes of the event itself and the &i, &M, &m and &w entries provide attributes of the @r entry. More than one @r entry can be provided and each can have its own set of &-key entries.

  • @-keys can generally be entered in any order provided that @s is entered before any options that require it. &-keys can also be entered in any order.

  • The prompts provide "just in time" information relevant to the entry you are typing and need only be consulted if you are uncertain about your entry. With a little experience, most reminders can be completed without a glance at the prompt.

fuzzy parsing of datetimes

Whenever etm expects a datetime entry as, for example, when you are entering an @s starting datetime, it applies fuzzy parsing to your entry. Suppose it is Dec 17 2019, your computer is in the Eastern timezone and you have just entered @s for your lunch event

    starting date or datetime

    * lunch @s_

If you enter 1, etm will interpret it as 1am on the current date in the local timezone.

    Tue Dec 17 2019 1:00am EST

    * lunch @s 1_

Adding p changes the interpretation to 1pm.

    Tue Dec 17 2019 1:00pm EST

    * lunch @s 1p_

Now start adding 'fri' by appending an 'f'.

    '1p f' is incomplete or invalid

    * lunch @s 1p f_

and etm will complain that the entry is now either incomplete or invalid. Add the remaining 'ri'.

    Fri Dec 20 2019 1:00pm EST

    * lunch @s 1p fri_

and etm understands that you want the coming Friday, Dec 20. Now suppose that you will be in California on Friday and you want Pacific, not Eastern time. Then add an entry for @z.

    local datetime: Fri Dec 20 2019 4:00pm EST

    * lunch @s 1p fri_ @z US/Pacific

Note that local datetime is now prepended to the result which is still displayed in the local timezone, since that is the location of your computer, but the time has changed to 4:00pm EST which, of course, is the same time as 1:00pm PST. etm always displays times in the current local time zone. When you save this item, the time will be converted to universal time and the @z US/Pacific will be deleted - once the time has been converted, the original timezone is no longer relevant.

What if you had entered the 'fri' first?

    Fri Dec 20 2019

    * lunch @s fri

Here etm supposes again that you want the coming Friday, Dec 20 but without a time being specified, it is interpreted as a date rather than a datetime and thus without a timezone. Add the 1p and timezone the result will be the same as before.

As a final illustration, suppose you want to be reminded of lunch at 1pm in whatever timezone you happen to be, i.e., you want the datetime to float to the local timezone. Add the @z.

    a timezone entry such as 'US/Eastern' or 'Europe/Paris'
    or 'float' to specify a naive/floating datetime

    * lunch @s fri 1p @z

and note that supplying 'float' as the timezone does the trick.

    Fri Dec 20 2019 1:00pm

    * lunch @s fri 1p_ @z float

The datetime now appears without either the 'local datetime' prefix or the 'EST' suffix which reveals that it is a naive/floating datetime.

The assumption here is that when a user enters a date, a date is what the user wants. When both a date and time are given, what the user wants is a datetime and, most probably, one based on the local timezone. Less probably, one based on a different timezone and that requires the addition of the @z and timezone. Still less probably, one that floats and this requires the addition of the @z and 'float'.

Note: When entering dates, ., - or / can be used to separate the components. The interpretation of the date depends upon the 'dayfirst' and 'yearfirst' settings in the etm configuration file - see configuration for details about these settings. For example, with both dayfirst and yearfirst false,

6/1 => June 1 in the current year

but after changing dayfirst to true

6/1 => January 6 in the current year

relative datetimes

Relative datetimes can be entered using period strings either instead of or in addition to datetimes using the format:

[datetime] [period string]

Either the datetime or the period string is required but both are allowed. The period string must begin with either a plus or a minus sign.

If the datetime is omitted and

  • the period string involves d (days), w (weeks) or M (months), then the assumed datetime is the current date,
  • the period string involves only h (hours) or m (minutes), then the assumed datetime is the current time

Examples supposing it is currently 1:20pm on July 15:

+1h30m     => 2:50pm on July 15
+3d        => July 18
-3d        => July 12
8a         => 8am on July 15
8a +1h30m  => 9:30am on July 15
8a +3d     => 8am on July 18

tab completion

Some optional attributes that can be used in reminders can be tedious to enter either because they are lengthy or because they are used often and typographical errors would introduce inconsistencies. These attributes include the following:

  • @g: goto url or filepath. In the latter case, the system default application is used to open the filepath, e.g., on my macbook the system default for files ending in .py is to open them in MacVim. In the former case, the url will be opened in the default browser. In either case, a typo would likely render the link useless.
  • @l: location/context. In the do next view, location is used to organize undated tasks. Using @l errands in one task and @l erands in another would cause these tasks to be grouped separately.
  • @i: index - colon delimited string. In both the index and the used time views, the index entry is used to group and aggregate entries. Using @i jsmith:project a in one entry and @i jsmth:project a in another would cause these these entries to be grouped and aggregated separately.
  • @n: attendee. string using "[name:] address" format. If "address" begins with exactly 10 digits followed by an "@" it is treated as a mobile phone number. Otherwise it is treated as an email address. The optional "name:" can be used to facilitate autocompletion. Attendee entries are used in both email and text alerts.

Other tracked attributes include @c (calendar) and @t (tag).

For these attributes, etm keeps a record of each usage. I.e., for each reminder in your etm database, if the reminder includes a tracked attribute that has not already been included in the list of completions, then that attribute will be added. E.g., if you create an reminder containing

    @n joe:

and '@n joe:' is not already in the list of completions, then it would be added.

When you next create a reminder and enter @n, etm will pop up a sorted list of every item in the completion list that begins with @n. You can use the up and down cursor keys to move through this list or you can continue typing to limit the possible completions, e.g., after entering '@n joe' only those completions that start with '@n joe' would still be displayed. When you have selected the completion you want, press the space bar to insert the completion in your entry.


etm has several ways of viewing entries. These are listed below by the shortcut key used to activate the view. E.g., pressing a activates Agenda view shown below. In each of the views, the etm menus appear at the top and the status bar at the bottom. The latter displays the current datetime and the name of the view.


The display for each reminder shows the itemtype and summary column on the left followed by a flags column which displays a combination of letters from 'g' (goto), 'k' (connection), 'p' (pinned) and 't' (timer) reflecting the presence of these attributes in the reminder. This column is sometimes followed on the extreme right by another column whose contents depends on the view. E.g. in index and journal views the 'id' of the reminder is displayed while in history view, the last modification timestamp of the reminder is displayed.

  • a: Agenda: dated unfinished tasks and other reminders by year-week and week day
  • b: Busy: a graphical illustration of busy and conflicted times by year-week
  • c: Completed: finished tasks and jobs and used time entries by year-week and week day
  • d: Do Next: undated tasks grouped by location
  • f: Forthcoming: unfinished dated tasks and other dated reminders by next occurrence
  • h: History: all items by the latter of the modified or created datetimes in descending order, i.e., most recent first. Datetimes are displayed using a 5 character format where, e.g., 1:15pm today would be displayed as 13:15, November 7 of the current year as 11/17 and January 15 of 2012 as 12.01.
  • i: Index: all items grouped hierarchically by index entry
  • j: Journal: journal entries grouped hierarchically by index entry
  • k: Konnection: items with @k konnection links either to or from the selected item.
  • l: Location: items grouped hierarchically by location entry
  • m: Timers: items with timers.
  • p: Pinned: items whose pin status is on.
  • q: Query: items matching a user specified query. Enter ? for query usage.
  • r: Review: undated tasks sorted by the time since the task was last modified with the most recently modified last.
  • t: Tags: all items with @t tag entries grouped by tag
  • u: Used Time: all items with @u used time entries grouped by month and hierarchically by index
  • U: Used Summary: used time aggregates grouped by month and hierarchically by index
  • y: Yearly Planning Calendar: compact monthly calendar by half year.

Weekly Views

The weekly agenda, busy and completed views display one week at a time and are synchronized so that all three views always display the same week. Left or right cursor keys go backward or forward a week at a time and the pressing the space bar jumps to the week containing the current day. You can also press "j" and enter a date to jump to the week containing the date.

In both agenda and completed views, only days with scheduled reminders are listed. If nothing is scheduled for the entire week, then "Nothing scheduled" is displayed.

The normal agenda listing for a week day:

  • all day events (events with dates as @s entries)
  • datetime items (reminders with datetimes as @s entries) by time
  • all day tasks (tasks with dates as @s entries)
  • all day journal enties (journal entries with dates as @s entries)

And, on the current day only:

  • inbox items
  • < pastdue warnings in descending order of the number of days past due
  • > beginby warnings in ascending order of the number of days remaining

Timer view

This view lists all reminders with associated timers sorted by the elapsed time since the timer's state was last changed. The display for each reminder shows the itemtype and summary, any applicable flags and, in the right hand column, the elapsed time and state of the associated timer.

The sort order assures that the reminder with the active timer will always be at the top of the list and followed by the reminders with the most recently modified timers. This makes it easy to switch back and forth between recent timers.

Review View

Displays a list of the summaries and location/contexts (@l entries) of undated and unfinished tasks sorted by their (last) modified timestamp and grouped by the number of weeks since last modified.

This view is used for a periodic review of such "todos" with the goal of not letting them 'slip through the cracks'. Either editing a task or pressing "^U" with the task selected resets the modified timestamp to the current time and thus moves the task to the "this week" group at the bottom of the list.

A reasonable work flow would be to open this view once every week or so and examine tasks more than a week "old", editing them when necessary and otherwise updating the modified timestamp using "V" so that all tasks are kept within the 'last week' or 'this week' groups.

Pinned View

Items that have been "pinned" (have their pin status toggled on) are displayed in pinned view grouped by itemtype and sorted by the created/modified datetime.

This view can be used to flag items that need attention in your daily workflow in the same way that you might flag email in your inbox that you want to handle first. Just as you might sort your inbox to move the flagged items to the top, you can switch to pinned view to see just the reminders that need attention. And just as you would remove the flag when you finish with an email, you can unpin the reminder that no longer needs attention and thus remove it from the pinned view.

The pinned status of items is retained so long as etm is active but cleared when etm is restarted.

Konnected View

Items with @k konnection links either to or from other items are displayed with a 'k' in the flags column of normal views. When such an item is selected and 'k' is pressed this view displays all reminders konnected to the selected reminder, organized as follows:

reminders with links to the selection : the list of items with @k entries which include the id of the selected item

the selection : the selected item

reminders with links from the selection : the list of items whose ids are included in the @k entries of the selected item

Used Time Views

The used time and used time summary views are bound to u and U respectively. They report @u (used time) entries in your reminders grouped by year-month and then heirarchially by @i entries. I have a file of reminders with @i and @u entries such as

	* Modi ut sit sed amet sit @s 2019-11-11 10:00am @e
	@u 58m: 2019-11-11 10:58am @u 34m: 2019-11-11 10:34am
	@i client A/project a1/correspondence
	@d Aliquam non sed aliquam eius tempora quisquam dolorem.
	Neque quiquia labore tempora magnam. Quiquia tempora
	porro est ut. Ut tempora sed non ut eius neque porro.
	Sed quaerat consectetur dolor sit.

and the used time view for November begins with

	November 2019
	  client A
		project a1
			* Modi ut sit sed amet sit: 1.6h Nov 11
			% Amet modi neque eius adipisci: 2.7h Nov 27
			* Quisquam quiquia velit non: 2.0h Nov 19
		project a2
			* Consectetur voluptatem dolorem: 1.0h Nov 6
			* Quaerat etincidunt sed non: 0.9h Nov 13
			* Consectetur eius est adipisci: 0.5h Nov 25
			% Magnam labore etincidunt: 1.8h Nov 28
			% Adipisci dolor labore quiquia: 0.9h Nov 7
			% Adipisci eius velit porro: 1.4h Nov 14
			* Non modi non velit eius: 1.0h Nov 20
	  client B
		project b1
			* Dolor neque velit dolorem: 0.4h Nov 22
			* Ipsum numquam porro consectetur: 0.8h Nov 15
			- Porro voluptatem aliquam: 1.0h Nov 12
			% Amet ut dolor velit aliquam: 1.9h Nov 13
			* Quisquam labore ut sit aliquam: 0.7h Nov 5
			* Quiquia ut quisquam sit: 1.5h Nov 12
			% Adipisci amet modi sed eius: 2.6h Nov 15
			- Velit dolor quiquia etincidunt: 1.9h Nov 15

Note that the display is by month and, within the month, heirarchially by index entry and reminder. Note also, that the reported times are aggregates of all @u entries in the reminder. The first reminder, for example, has 2 such entries:

	@u 58m: 2019-11-11 10:58am @u 34m: 2019-11-11 10:34am

Because of

	usedtime_minutes: 6

in cfg.yaml, each @u timeperiod is first rounded up to the next 6 minutes and then added. Thus 58m becomes 1h, 34m becomes 36m and the sum, 1h36m, is reported in hours and tenths as 1.6h.

The reminder lines are similar to those in other views. With, e.g.,

		* Modi ut sit sed amet sit: 1.6h Nov 11

selected, pressing return would display the item's details, pressing E would open it for editing and so forth.

The used time summary view for the same month begins with:

	November 2019: 44.4h
	   client A: 13.8h
		  project a1: 6.3h
			 correspondence: 4.3h
			 research: 2.0h
		  project a2: 7.5h
			 correspondence: 4.2h
			 meeting: 2.3h
			 research: 1.0h
	   client B: 16.6h
		  project b1: 10.8h
			 correspondence: 0.4h
			 meeting: 0.8h
			 phone: 2.9h
			 research: 6.7h
		  project b2: 3.9h
			 meeting: 1.8h
			 research: 2.1h
		  project b3: 1.9h
			 correspondence: 1.9h

This view omits the reminder lines and aggregates the used times hierarchically by index entry.

As with other dated views, the left and right cursor keys go backwards and forwards a month at a time and the space bar returns to the current month. Also, pressing ^C copies the contents of the view to the system clipboard.

Query View

In query view an entry line at the bottom of the screen is used to submit queries to your data store of reminders. For example, press q to open query view, enter

	includes summary waldo

and press return to see a list of reminders in which either the summary or the @d element includes a match for the case-insensitive regular expression "waldo". Each line of the display contains the item type, the summary and the document id of the matching reminder. As another example

	exists u and ~exists i

would display reminders with an @u element but not an @i element.

Simple queries of this type produce a list of matching items with the itemtype, summary and id displayed and sorted by id, i.e., by the order created. It is also possible to create more complex queries in which the output is displayed heirarchially using a format determined by the query parameters. Enter ? or help at the prompt to get this detailed usage information:

Simple queries

Return a list of items displaying the itemtype, summary and id, and sorted by id, (order created) using commands with the format:

command field [args]

where "field" is either 'itemtype', 'summary' or one of the '@-keys' such as 'l' or 's', and "command" is one of those listed below (see Simple Query Examples below for examples):

  • begins field RGX: return items in which the value of field begins with a match for the case insensitve regular expression RGX.

  • includes LST RGX: return items in which the value of one of the fields in LST includes a match for the case insensitive regular expression RGX. (LST contains all but the last, RGX, argument.)

  • equals field VAL: return items in which the value of field == VAL

  • more field VAL: return items in which the value of field >= VAL. The value of field and VAL must be comparable, i.e., both strings or both numbers

  • less field VAL: return items in which the value of field <= VAL. The value of field and VAL must be comparable, i.e., both strings or both numbers

  • exists field: return items in which field exists

  • any field LST: return items in which the value of field is a list and at least one element of field is an element of LST

  • all field LST: return items in which the value of field is a list and the elements of field contain all the elements of LST

  • one field LST: return items in which the value of field is one of the elements of LST

  • info ID: return the details of the item whose document id equals the integer ID

  • dt field EXP: return items in which the value of field is a date if EXP = '? date' or a datetime if EXP = '? time'. Else if EXP begins with '>', '=' or '<' followed by a string following the format 'yyyy-mm-dd-HH-MM' then return items where the datetime of the field value bears the specified relation to the string, with hours and minutes ignored when the value of field is a date. E.g.,

      dt s < 2020-1-17

    would return items with @s date/times whose year <= 2020, month <= 1 and month day <= 17.


      dt s ? date and equals itemtype *

    would return events with @s date/times that are dates, i.e., all day events or occasions.

Enter the command at the 'query:' prompt and press 'enter' to submit the query and display the results. Press 'q' to reopen the entry area to submit another query. Use up and down cursor keys to choose from the command history (see Command History below), submit '?' or 'help' to show this display, submit 'l' to see a list of stored queries (see Saved queries below) or submit 'quit', 'exit' or nothing at all, '', to close the entry area and return to the previous display.

Simple query examples

Find items where the summary includes a match for "waldo":

query: includes summary waldo

Precede a command with ~ to negate it. E.g., find reminders where the summary does not include a match for "waldo":

query: ~includes summary waldo

To enter a list of values for "arg", simply separate the components with spaces. E.g.,

query: includes summary d waldo

would return items in which either the summary or d (@d description) contains a match for "waldo". Similarly

query: all t blue green

would return items with both blue and green tags or

query: any t blue green

would return items with either a blue or a green tag. As a last example

query: one itemtype - *

would return items in which the item type is either '-' or '*', i.e., either a task or an event.

Conversely, to enter a regex with a space and avoid its being interpreted as a list, replace the space with '\s'. E.g.,

query: matches i john\sdoe

would return items with '@i' (index) entries such as "John Doe/...".

Note: in the world of regular expressions '\s' matches any white space character, including a space.

Components can be joined the using "or" or "and". E.g., find reminders where the summary entry contains a match for "waldo" but the @d (description) entry does not:

query: includes summary waldo and ~includes d waldo

Archive queries

Queries, by default, search the items table in the etm database. You can preceed any query with 'a ' (the letter 'a' followed by a space), to search the archive table instead. E.g.,

query: a includes summary waldo or includes d waldo

will search the archive table for reminders with matches for 'waldo' in the summary or in the description.

Queries beginning with 'a ' are, in fact, the only way to see archived items from within etm itself.

Update queries

Queries can not only locate reminders but also update them. The update commands act on items returned by a query. E.g., this query

query: includes i john\sdoe | replace i john\sdoe

can be regarded as taking the reminders whose index entry, 'i', includes a match for 'john doe' and feeding them to the 'replace' command which then takes the index entry of each reminder and replaces each match for 'john doe' with 'Jane Doe'.

All of the update commands work the same way: a query that returns a list of reminders is followed by a pipe symbol surrounded by spaces, ' | ', and an update command with its arguments. Here is the complete list of these update commands:

  • archive: if the items table is active, move the reminders from the items table to the archive table, else vice versa

  • remove: remove the reminders from the database

  • replace field RGX VAL: in the value of 'field', replace matches for RGX with VAL. If the value of 'field' is a list, do this for each element of the list. Remember to replace spaces in in both RGX and VAL with '\s'

  • delete field: if the reminder has an entry for 'field', delete the entry

  • set field VAL: set the value of 'field' to VAL overriding the current value if it exists

  • provide field VAL: set the value of 'field' to VAL if there is no existing entry for 'field'

  • attach field VAL: if the value of 'field' is intended to be a list and if VAL is not in this list, then add VAL to the list, creating the entry if necessary

  • detach field VAL: if the value of 'field' is a list and this list contains VAL, then remove VAL from the list

The recommended workflow for updating reminders is first to perfect the query to be certain that it lists just the items you want to update. Then press 'q' and the 'up' cursor key to restore the previous query, add ' | ' and the update command you want with its arguments.

WARNING: Since the results may not be reversible, consider backing up your 'db.json' database before using update commands. This simple query, e.g., would PERMANENTLY DELETE ALL YOUR REMINDERS:

query exists itemtype | remove

Complex queries

Return a formatted, heirarchial display of items. Both the format and the items displayed are determined by the type of the query and the arguments provided. Since these queries can group and sort by date/times, these queries must begin by specifying which of the possible datetimes to use. There are four types of datetime specifications:

  • u: sort and group by datetimes in '@u' (used time) entries. Also report aggregates of times spent in these entries. Only items with '@u' entries will be reported.
  • s: sort and group by datetimes in '@f' entries in finished tasks and otherwise by '@s' entries. Only items with '@f' or '@s' entries will be reported.
  • c: sort and group by the 'created' datetime. All items will be reported.
  • m: sort and group by the 'modified' datetime if given else by the 'created' datetime. All items will be reported.

Complex queries follow the datetime specifier with a required group/sort specification consisting of a semicolon separated list with at least one of the following components:

  • index specification such as i, i[1:2] or i[1:]

    E.g. for an item with index entry '@i A/B/C': i = ['A','B','C'] i[0] = 'A' i[1] = 'B' i[2] = 'C' i[3] => error, list index out of range i[0:] = ['A','B','C'] i[:1] = ['A'] i[1:] = ['B','C'] i[1:2] = ['B'] i[:2] = ['A','B'] i[2:] = ['C'] i[3:4] = i[3:] = []

    Note: using slices such as i[1:2] rather than i[1] avoids 'list index out of range errors' for index entries missing the indicated position and is strongly recommended.

    When an index specification returns an empty list, '' is used for the missing entry. Items without an '@i' entry are given a default entry of '' and included by default. Include 'exists i' in '-q' (discussed below) to overrule this default.

  • field specification: l: location c: calendar

    Note: items without the specified field are given a default entry of '~' and included by default. Include 'exists l' or 'exists c' in '-q' (discussed below) to overrule these defaults.

  • date specification:

    • year:
      • YY: 2-digit year
      • YYYY: 4-digit year
    • month:
      • M: month: 1 - 12
      • MM: month: 01 - 12
      • MMM: locale abbreviated month name: Jan - Dec
      • MMMM: locale month name: January - December
    • week: (examples based on 2020 iso week number 3):
      • W: week number: 3
      • WW: month days interval for week: Jan 13 - 19
      • WWW: interval and year: Jan 13 - 19, 2020
      • WWWW: interval, year and week number: Jan 13 - 19, 2020 #3
    • day:
      • D: month day: 1 - 31
      • DD: month day: 01 - 31
      • ddd: locale abbreviated week day: Mon - Sun
      • dddd: locale week day: Monday - Sunday

Note: when a date specification is given, the datetime used depends upon the report type.

  • u: the value of the datetime component of the @u entry. Items without @u entries are omitted.
  • s: the value of @f when it exists and, otherwise, the value of @s. Items lacking both @f and @s entries are omitted.
  • c: the created datetime.
  • m: the modified datetime if it exists, else the created datetime.


query: u i[:1]; MMM YYYY; i[1:]; ddd D

would create a usedtime query grouped (and sorted) by the first component of the index entry, the month and year, the remaining components of the index entry and finally the month day.

Sorting note: Specifications using weeks are all sorted and grouped by by (YYYY, W). Specifications involving months are all sorted by (YYYY, M). Specifications involving days are all sorted by (D).

The group/sort specification can be followed, optionally, by any of the following:

-b begin date/datetime: omit items with earlier datetimes

-e end date/datetime: omit items with later datetimes

-q query: exclude items not satisfying this simple query. Anything that could be used in a simple query described above could be used here. E.g., "-q exists f" would display only items with an "@f" entry, i.e., finished tasks. Similarly "-q equals itemtype - and ~exists f" would limit the display to unfinished tasks.

-a append: append the contents of this comma separated list of @key characters to the formatted output. E.g., "-a d, l" would append the item description and location to the display of each item.

Note: -b and -e accept shortcuts:

  • daybeg: 12am on the current day
  • dayend: 12am on the following day
  • weekbeg: 12am on Monday of the current week
  • weekend: 12am on Monday of the following week
  • monthbeg: 12am on the 1st of the current month
  • monthend: 12am on the 1st of the following month

and can be combined with period strings using M (month), w (week), d (day), h (hour) and m (minute). E.g.:

  • weekbeg - 1w (the beginning of the previous week)
  • monthend + 1M (the end of the following month)

Command History

Any query entered at the 'query:' prompt and submitted by pressing 'Enter' is added to the command history. These queries are kept as long as 'etm' is running and can be accessed using the up and down cursor keys in the query field. This means you can enter a query, check the result, press 'q' to reopen the query prompt, press the up cursor and you will have your previous query ready to modify and submit again. It is also possible to keep a permanent list of queries accessible by shortcuts. See 'Saved Queries' below.

Saved Queries

Commonly used queries can be specified in the "queries" section of cfg.yaml in your etm home directory along with shortcuts for their use. E.g. with the default entry

    # unfinished tasks by location and modified datetime
    md: m l -q equals itemtype - and ~exists f
    # usedtimes by i[:1], month and i[1:2] with d
    ut: u i[:1]; MMM YYYY; i[1:2] -a d
    # finish/start by i[:1], month and i[1:2] with u and d
    st: s i[:1]; MMM YYYY; i[1:2] -a u, d
    # items with an "@u" but missing the needed "@i"
    mi: exists u and ~exists i


query: ut

and pressing 'enter' would result in the 'ut' being replaced by its corresponding value to give

query: u i[:1]; MMM YYYY; i[1:2] -a d

This query can now be submitted as is or first edited to add, say, -b and -e options and then submitted. As with other queries, the submitted form of the query is added to the command history.


query: l

to display a list of the saved keys and values.

As with other etm views, in query view you can enter / or ? to search incrementally forward or backward, resepectively, or press Ctrl-C to copy the view to the system clipboard or select a reminder and press Enter to display its details, press E to edit it and so forth.

Common Features

While the views differ in many respects, they also share some common aspects:

  • Press N in any view to create a new item.
  • Select a reminder by clicking on it or by using the up and down cursor keys to move the cursor to the line displaying the reminder.
  • With a reminder selected, press return to toggle displaying the details.
  • Movement
    • Press page up or page down to shift the display a page at a time.
    • Press “l” (lower case L) and enter a number to move the cursor to a particular line number.
    • In the weekly views a), b) and c), press “j” and enter a date to display the week containing that date.
    • In the dated views a), b), c), u), U) and y), press the right or left cursor keys to go to the next or previous period, respectively, and the space bar to return to the current period.
  • Search.
    • Press “/“ (or "?") and enter an expression to search the view forward (or backward) for a row whose content contains a case-insensitive match for the expression.
    • While entering the search expression, push the up or down cursor keys to change the direction of search.
    • After entering the search expression, press “n” to search (cyclically) for other matches in the direction specified.
    • Once a search is initiated, it remains active in all views with matches highlighted. To remove the highlighting, search for something unlikely to be matched, e.g., 3 consecutive commas.


Pressing F1 toggles the etm menu display - opening it if it is closed and closing it if it is open. There are four menu tabs labeled etm, view, editor and selected with the options listed below:

	F1) activate/close menu
    F2) about etm
    F3) system info
    F4) check for updates
    F5) import file
    F6) datetime calculator
    F7) configuration settings
    F8) help
    ^q) quit
    a) agenda
    b) busy
    c) completed
    d) do next
    f) forthcoming
    h) history
    i) index
    j) journal
    l) location
    p) pinned
    q) query
	r) review
    t) tags
    u) used time
    U) used time summary
    s) scheduled alerts for today
    y) half yearly calendar
    /) search forward
    ?) search backward
    n) next incrementally in search
    ^l) prompt for and jump to line number
    ^p) jump to next pinned item
    ^c) copy active view to clipboars
    J) jump to date in a), b) and c)
    right) next in a), b), c), u), U) and y)
    left) previous in a), b), c), u), U) and y)
    space) current in a), b), c), u), U) and y)
    N) create new item
    ^s) save changes & close
    ^g) test goto link
    ^r) show repetitions
    ^z) discard changes and close
    Enter) toggle showing details
    E) edit
    C) edit copy
    D) delete
    F) finish
    P) toggle pin
    R) reschedule
    S) schedule new
    g) open goto link
    k) show konnections
    ^r) show repetitions
	^u) update last modified
    ^x) toggle archived status
    T) change timer to next state
    TR) record usedtime and delete timer
    TD) delete timer
    TT) toggle paused/running for active timer

etm menu notes

As with the other menus, each entry is preceeded by its shortcut, e.g., F2 for about etm or ^q (control and q simultaneously) to quit.

Many of the entries are obvious but a few deserve comment.

  • import file supports importing three different file types. A file ending with .json is expected to be one exported from etm 3.2.x. A file ending with .text is expected to be a text file with lines corresponding to etm 4.x reminders. A file ending with .ics is expected to be a file in iCalendar format.

    If a file named inbasket.text exists in the root of the etmdir directory, it will be offered as the default for importing. When etm detects the presence of this file, a circled-i symbol, 𝕚, will be displayed in the right-hand end of the status bar.

    Any file located in the root of the etmdir will automatically be removed after it is imported to avoid duplications.

  • datetime calculator processes an expression of the form x [+-] y where x is a datetime and y is either a timeperiod with + or a datetime or a timeperiod with -. As an example, suppose you have the arrival time in Paris of a flight and the departure time from Raleigh/Durham and you would like to determine the flight time. Entering

      	7:45a 4/7 Europe/Paris - 5:30p 4/6 US/Eastern


      	8 hours 15 minutes
  • configuration settings opens the file cfg.yaml using the default text editor for your operating system. Note that any changes you make to this file will not take effect until you close and reopen etm.

  • help opens the etm documentation on google pages using the default web browser for your system. This is the most user friendly source for the documentation because it begins with a table of contents whose elements are active links to the relevant sections. It is updated with every commit so it is always the most recent version available.

view menu notes

The view menu provides access to all the etm views with the shortcut keys for these views. E.g., press a to open agenda view.

The entries here are pretty obvious and the views themselves are descibed elsewhere.

editor menu notes

It is worth noting here that when you are editing an item, ^s (control and s) saves any changes you have made and closes the editor. ^z (control and z), on the other hand, closes the editor without saving any changes but, if there are changes, asks for confirmation that this is what you want.

An options setting, 'vi_mode' determines which keybindings are used in the entry buffer when editing. If 'vi_mode' is true, then vi-style bindings are used and, otherwise, the default emacs-style bindings are used. The status bar describes the current mode, e.g., 'vi: insert' with a '+' appended if the contents have been modified.

selected menu notes

Options in the selected menu are only relevant when a reminder has been selected in one of the views.

Several options here deserve comment.

  • delete will prompt for comfirmation and, if the selected item is repeating, for whether to delete only the selected instance or the item itself, i.e., all instances.

  • finish applies only to unfinished tasks and will prompt for a datetime to use in creating an @f entry for the task with the current datetime as the default. Press

  • reschedule will prompt for a datetime. If the reminder is repeating, the provided datetime will replace the datetime of the selected instance. Otherwise it will be used either to replace the current value of @s or, if there is no @s entry, to create one.

  • schedule new will prompt for a datetime and add that instance to any other instances of the reminder.

  • open goto will use the system default application to open the file path or url specified in the selected reminders @g entry. Items with goto links are displayed with a 'g' in the flags column of normal views.

  • toggle pin toggles the pin status of an item between off and on. Items for which the pin status are displayed with a 'p', in the flags column of normal views and are also displayed in pinned view.

  • show repetitions pops up a display showing illustrative repetitions if the item is repeating.

  • toggle archived status moves the selected reminder from the items table if it is active to the archive table and vice versa if the archive table is active.

  • change timer to next state.

    • these are the states for the timer associated with a reminder and the relevant status of any other timers:

      • n (no timer is associated with this reminder)
        • n-: no other timer is active
        • n+: another timer is active
      • i (a timer is associated with this reminder and it is inactive)
        • i-: no other timer is active
        • i+: another timer is active
      • r-: running (a timer is associated with this reminder and it is running - there cannot be another active timer)
      • p-: paused (a timer is associated with this reminder and it is paused - there cannot be another active timer)

      These are the "next" state transitions associated with this action:

      • n+i+
      • n-r-
      • i-r-
      • i+r-
      • r-p-
      • p-r-
    • Here are the details:

      • If no timer is currently associated with the reminder, one will be created. If another timer is active, the new timer will be inactive, otherwise it will be running.
      • If an inactive timer is already associated with the reminder. Its state will be changed to running. If another timer is active, that timer will be changed to inactive and, if running, its elapsed time will be incremented by the period it has been running.
      • If an active timer is already associated with the reminder, its state will be toggled between paused and running. With a transition from running to paused, the elapsed time for the timer in increased by the period since the timer's state changed to running.

      Additionally, if there is an active timer, its state, either r (running) or p (paused), is displayed in the etm status bar together with the relevant time period, either the total elapsed time or how long the timer has been paused. If there are inactive timers with non-zero elapsed times, the total of such elapsed times will also be reported in the status bar. E.g., this display in the status bar

        r:4m + i:13m

      would mean that the active timer is running with 4 minutes of unrecorded elapsed time and that there is an additional 13 minutes of unrecorded elapsed time in inactive timers.

  • record used time with a reminder selected, will:

    • If a timer is not currently associated with the reminder, prompt for a timeperiod and ending time to use in adding a usedtime entry to the reminder.
    • If a timer is currently associated with the reminder, use its elapsed time and the last moment it was running as the defaults in the prompt for a usedtime timeperiod and ending time. If the entry is saved, the elapsed time for the timer will be reset to zero and, additionally, if the timer is currently running, its status will be changed to paused.

    Important: timer data is stored in memory and will be lost if etm is stopped. Be sure to record usedtime entries for timers with elapsed times before quitting etm. The status bar will always indicate if there is unrecorded elapsed time.


For use in a virtual environment

Setting up a virtual environment for etm is recommended for new users. The steps for OS/X or linux are illustrated below. For details see python-virtual-environments-a-primer.

Open a terminal and begin by creating a new directory/folder for the virtual environment, say etm-pypi in your home directory:

    $ mkdir ~/etm-pypi
    $ cd ~/etm-pypi

Now continue by creating the virtual environment (python >= 3.7.4 is required for etm):

    $ python3 -m venv env

After a few seconds you will have an ./env directory. Now activate the virtual environment:

    $ source env/bin/activate

The prompt will now change to something containing (env) to indicate that the virtual environment is active. Updating pip is now recommended:

    (env) $ pip install -U pip

Note that this invokes ./env/bin/pip. Once this is finished, use pip to install etm:

    (env) $ pip install -U etm-dgraham

This will install etm and all its requirements in


and will also install an executable called etm in ./env/bin.You can then start etm using

    (env) $ etm <path to home>

Details about the home directory are in usage.

For use system wide

If your system allows you to run sudo and you want general access system wide, then you could instead install etm using

$ sudo -H python3.x -m pip install -U etm-dgraham

replacing the 3.x with the verion of python you want to use, e.g., 3.7. This would put both etm and etm+ in your path (in the bin directory for python3.7).


  • This same command would be used to update etm to the latest version.

  • You may or may not need the '-H' argument for sudo. Here is the relevant section from the sudo man page:

      -H, --set-home
                  Request that the security policy set the
                  HOME environment variable to the home
                  directory specified by the target user's
                  password database entry.  Depending on
                  the policy, this may be the default
  • Invoking pip through python in this way forces the use of the pip that belongs to python3.7.

You can then open any terminal and start etm using

$ etm <path to home>


Terminal size and color scheme

The suggested terminal size for etm is 60 (columns) by 32 or more (rows). The default color scheme is best with a dark terminal background. A scheme better suited to light backgrounds can be set using style: light in cfg.yaml in your home directory. Some of the etm display may not be visible unless style is set correctly for your display.

The size of the terminal is used when etm starts to set various display options so changing the terminal size, especially reducing the width, is best avoided once etm is running.

Home directory

Before you start etm, think about where you would like to keep your personal data and configuration files. This will be your etm home directory. The default is to use whatever directory you're in when you start etm as your etm home directory. If you start etm in your virtual environment directory then the default will be to use that as your home directory as well. If this is not what you want, you can just give the path for whatever directory you would like to use when you start etm.

$ etm <path to home>

Finally, if there is an environmental variable, ETMHOME, set to this path then you can just enter

$ etm

and etm will use ETMHOME as its home directory.

Home directory considerations:

  • If more than one person will be using etm on the same computer, you might want to have different home directories for each user.
  • If you want to use etm on more than one computer and use Dropbox, you might want to use ~/Dropbox/etm to have access on each of your computers.
  • If you want to separate personal and professional reminders, you could use different home directories for each. You can run two instances of etm simultaneously, one for each directory, and have access to both at the same time.
  • You can have multiple instances of etm running with the same home directory, provided that you treat all instances save one as read only, i.e., only make changes in one instance.

Whatever home directory you choose, running etm for the first time will add the following to that directory.

    etm home directory/

Here cfg.yaml is your user configuration file and db.json contains all your etm reminders. The folders backups/ contains the 7 most recent daily backups of your db.json and cfg.yaml files. The folder logs contains the current etm.log file and the 7 most recent daily backups. Note that backup files are only created when the relevant file has been modified since the last backup.

The file cfg.yaml can be edited and the options are documented in the file. See configuration for details.

Using etm+

An added bonus of setting ETMHOME to the path of your home directory is the possibility of using the etm+ shortcut for creating reminders. E.g., entering

$ etm+ '* lunch with Peter @s fri 12p'

would append the line * lunch with Peter @s fri 12p to the file inbasket.textin ETMHOME, creating the file if necessary. etm+ also accepts input piped to it so that

$ echo '* lunch with Peter @s fri 12p' | etm+

would produce exactly the same result.

Important : The single quotes are necessary to keep the shell from expanding the "*" into the names of all the files in the current working directory and other such mischief.

etm checks once every minute for the presence of a file named inbasket.textin ETMHOME and, if found, will display an inbasket character, ⓘ , at the right end of its status bar reminding you that inbasket items are available for importing. Just press F5 in etm to import the reminders from this file and, on successful completion, automatically remove the file.

Note finally that etm+ will accept quick notes which are not themselves valid etm reminders such as

$ etm+ '123 456-7890 Peter'

This would result in the valid reminder

! 123 456-7890 Peter @t etm+

being appended to inbasket.text - note the added type character, ! and the tag, @t etm+.

The addition of the typechar '!' means that after importing the reminder will appear as an 'inbox' item. These are highlighted in the list for the current day in agenda view, reminding you that they require futher attention. You can thus make quick notes without much thought and know that you will automatically be reminded to sort them out later. An added bonus is that when editing such entries in etm itself, all its completion, fuzzy parsing and verification features are available.


From a virtual environment

If you should ever want to deinstall etm, first deactivate the virtual environment, if necessary, by changing to the virtual environment directory and entering

    (env) $ deactivate

You can now simply delete the virtual environment directory and, if you have additional home directories, delete each of them. One of the many advantages of the virtual environment is that these steps remove every trace.

From a system wide installation

To remove etm installed into, say, python3.7, run

$ sudo -H python3.7 pip uninstall etm-dgraham

This will remove etm from the python site-packages directory and the etm and etm+ executables from the python bin directory. Then remove any etm home directories that you have created.


Item Types


Type character: *

An event is something that happens at a particular date or datetime without any action from the user. Christmas, for example, is an event that happens whether or not the user does anything about it.

  • The @s entry is required and can be specified either as a date or as a datetime. It is interpreted as the starting date or datetime of the event.
  • If @s is a date, the event is regarded as an occasion or all-day event. Such occasions are displayed first on the relevant date.
  • If @s is a datetime, an @e entry is allowed and is interpreted as the extent or duration of the event - the end of the event is then given implicitly by starting datetime plus the extent and this period is treated as busy time.

Corresponds to VEVENT in the vcalendar specification.


Type character: -

A task is something that requires action from the user and lasts, so to speak, until the task is completed and marked finished. Filing a tax return, for example, is a task.

  • The @s entry is optional and, if given, is interpreted as the date or datetime at which the task is due.

    • Tasks with an @s datetime entry are regarded as pastdue after the datetime and are displayed in Agenda View on the relevant date according to the starting time.
    • Tasks with @s date entry are regarded as pastdue after the due date and are displayed in Agenda View on the due date after all items with datetimes.
    • Tasks that are pastdue are also displayed in Agenda View on the current date using the type character < with an indication of the number of days that the task is past due.
  • Tasks without an @s entry are to be completed when possible and are sometimes called todos. They are regarded as next items in the Getting Things Done terminology and are displayed in Do Next view grouped by @l (location/context).

  • Tasks with an @r (repeat) entry can have an @o (overdue) setting.

    • @o k: keep. Whenever completed, the next instance is due at the datetime specified in the recurrance rule even if that datetime has already passed. E.g. mortage payments to be made on the 1st of the month are due for each prior month in which they have not been made. With this option, many instances can be past due. This is the default when no @o entry is given.
    • @o r: reset. Whenever completed, the next instance is due at the first datetime from the recurrance rule that falls after the current datetime. E.g., getting a haircut every 14 days is due 14 days after the last haircut. With this option, at most one instance can be past due.
    • @o s: skip. Like 'keep' and 'reset' combined with the addition that pastdue instances are ignored. E.g., taking out the trash every Monday morning for pickup is due every Monday morning but, if a Monday passes without taking out the trash, the instance is better regarded as irrelevant than past due. With this option, an instance can never be past due.
  • Jobs

    • Tasks, both with and without @s entries can have component jobs using @j entries.

    • For tasks with an @s entry, jobs can have an &s entry to set the due date/datetime for the job. It is entered as a timeperiod relative to the starting datetime (+ before or - after) for the task. Zero minutes is the default when &s is not entered.

    • For tasks with an @s entry, jobs can also have &a, alert, and &b, beginning soon, notices. The entry for &a is given as a time period relative to &s (+ before or - after) and the entry for &b is a positive integer number of days before the starting date/time to begin displaying "beginning soon" notices. The entry for @s in the task becomes the default for &s in each job. E.g., with

        - beginning soon example @s 1/30/2018
            @j job A &s 5d &b 10
            @j job B &b 5

      Beginning soon notices would begin on Jan 15 for job A (due Jan 25) and on January 25 for job B (due Jan 30).

    • Prerequisites

      • Automatically assigned. Here it is supposed that jobs must be completed sequentially in the order in which they are listed. E.g., with

          - automatically assigned
              @j job A
              @j job B
              @j job C
              @j job D

        job A has no prerequisites but is a prerequisite for job B which, in turn, is a prerequisite for job C which, finally, is a prerequisite for job D.

        Auto assignment is done when the task is first saved by assigning 'a', 'b', 'c', ... as ids for the successive jobs with 'a' as a prerequisite for 'b', 'b' for 'c' and so forth. The task is then saved as

          - automatically assigned
              @j job A &i a
              @j job B &i b &p a
              @j job C &i c &p b
              @j job D &i d &p c

        and thereafter treated as if the ids and prerequisites had been manually assigned. Note that at most 26 jobs are possible with auto assignment.

        Once ids and prerequisites have been assigned, it is sometimes useful to be able to add a job here or there or change the order of existing jobs and have the ids and prerequisites generated again. To do this simply remove the &i and &p entries from the first job and save the task - all job ids and preqrequisites will automatically be reassigned.

      • Manually assigned. Job prequisites can also be assigned manually using entries for &i (id) and &p, (comma separated list of ids of immediate prequisites). E.g., with

          - manually assigned
              @j job a &i a
              @j job b &i b &p a
              @j job c &i c &p a
              @j job d &i d &p b, c

        job a has no prequisites but is a prerequisite for both job b and job c which are both prerequisites for job d. The order in which the jobs are listed is irrelevant in this case.

    • Tasks with jobs are displayed by job using a combination of the task and job summaries with a type character indicating the status of the job. E.g.,

        ✓ manually assigned [2/1/1]: job a
        - manually assigned [2/1/1]: job b
        - manually assigned [2/1/1]: job c
        + manually assigned [2/1/1]: job d

      would indicate that job a is finished, job b and job c are available (have no unfinished prerequistites) and that job d is waiting (has one or more unfinished prerequisties). The status indicator in square brackets indicates the numbers of available, waiting and finished jobs in the task, respectively.

  • An entry for @e can be given with or without an @s entry and is interpreted as the estimated time required to complete the task.

  • When a job is finished, the "done" datetime is recorded in an &f entry in the job. When the last job in a task is finished or when a task without jobs is finished a similar entry in the task itself using @f done. If there are jobs, then the &f entries are removed from the jobs.

  • Another step is taken for repeating tasks with as yet unfinished future repetitions. When the task or last job in the current repetition is completed, the @s entry is updated using the setting for @o to show the next due datetime and the @f entry is removed and appended to the list of completions in @h. A user configuration setting determines the number of most recent completion records retained for repeating tasks with 3 as the default.

  • When the last instance of a repeating task is finished, @f will contain the datetime of the last completion and @h the list of prior completions.

  • A task, repeating or not, will have an @f entry if and only if the task has been completed.

Corresponds to VTODO in the vcalendar specification.


Type character: %

A record of something that the user wants to remember. The userid and password for a website would be an example. A journal entry for vacation day is another example.

  • The @s is optional and, if given, is interpreted as the datetime to which the journal entry applies.
  • Journal entries without @s entries might be used to record personal information such as account numbers, recipies or other such information not associated with a particular datetime. They are displayed in the Journal view
  • Journal entries with @s entries associate the entry with the datetime given by @s. A vacation log entry, for example, might record the highlights of the day given by @s. They are displayed in Agenda view as well as Journal view.

Corresponds to VJOURNAL in the vcalendar specification.


Type character: !

An inbox item can be regarded as a task that is always due on the current date. E.g., you have created an event to remind you of a lunch meeting but need to confirm the time. Just record it using ! instead of * and the entry will appear highlighted in the agenda view on the current date until you confirm the starting time.

Corresponds to VTODO in the vcalendar specification.


These type characters are generated automatically by etm to display the status of reminders.

beginning soon

Type character: >

For unfinished tasks and other items with @b entries, when the starting date given by @s is within @b days of the current date, a warning that the item is beginning soon appears on the current date together with the item summary and the number of days remaining until the current date.

past due

Type character: <

When a task is past due, a warning that the task is past due appears on the current date together with the item summary and the number of days that the task is past due.


Type character: +

When a task job has one or more unfinished prerequisites, it is displayed using + rather than -.


Type character:

When a task or job is finished, it is displayed in the completed view on the date that it was finished using rather than -.



  • The term "list" means a comma and space separated list of elements, e.g.,

    @+ 2p Thu, 9p Mon, 11a Wed

  • Options displayed with an asterick can be used more than once, e.g.,

    @a 1h, 30m: t, v @a 1d: e

    @n joe: @n john:

    @t red @t green

@ keys

@ followed by a character from the list below and a value appropriate to the key is used to apply attributes to an item. E.g.,

        @s mon 9a

would specify the the starting datetime for the item is 9am on the Monday following the current date.

  • @+: include. list of datetimes to include
  • @-: exclude. list of datetimes to exclude
  • @a*: alert. list of + (before) or - (after) periods: list of commands
  • @b: beginby. integer (number of days before)
  • @c: calendar. string
  • @d: description. string
  • @e: extent. timeperiod
  • @f: finished. datetime
  • @g: goto. string (url or filepath)
  • @h: history. (for repeating tasks, a list of the most recent completion datetimes)
  • @i: index. forward slash delimited string. E.g., client/project/activity
  • @j*: job summary. string, optionally followed by job &key entries
  • @k*: doc_id. connect this reminder to the one corresponding to doc_id.
  • @l: location/context. string
  • @m: mask. string stored in obfuscated form
  • @n*: attendee. string using "[name:] address" format. If "address" begins with exactly 10 digits followed by an "@" it is treated as a mobile phone number. Otherwise it is treated as an email address. The optional "name:" can be used to facilitate autocompletion.
  • @o: overdue. character from (r) restart, (s) skip or (k) keep
  • @p: priority. integer from 0 (none), 1 (low), 2 (normal), 3 (high), 4 (urgent)
  • @r*: repetition frequency, a character from (y)early, (m)onthly, (w)eekly, (d)aily, (h)ourly or mi(n)utely, optionally followed by repetition &key entries
  • @s: start. date or datetime
  • @t*: tag. string
  • @u*: usedtime. string using "timeperiod spent: ending datetime" format
  • @x*: expansion. string
  • @z: timezone. string. A timezone specification, such as 'US/Eastern' or 'Europe/Paris' for aware datetimes or 'float', to indicate a naive or floating datetime. Datetime entries in the item are interpreted as belonging to the specified timezone when the entry is saved. The current timezone is the default when @z is not specified. Aware datetimes are converted to UTC (coordinated universal time) when stored and the @z entry, now irrelevant, is discarded.

& keys

& followed by a character from one of the lists below. These keys are only used with @j (job) and @r (repetition) entries.

For use with @j:

  • &s: start/due: period relative to @s entry (default 0m)
  • &a: alert: list of + (before) or - (after) periods relative to &s: list of cmd names from the users configuration file
  • &b: beginby. integer number of days before &s
  • &d: description. string
  • &e: extent. period
  • &f: finished. datetime
  • &l: location/context. string
  • &m: masked. string stored in obfuscated form
  • &i: job unique id. string
  • &p: prerequisites. list of ids of immediate prereqs
  • &u*: usedtime. string using "timeperiod spent: ending datetime" format

For use with @r:

  • &c: count. integer number of repetitions
  • &E: easter. number of days before (-), on (0) or after (+) Easter
  • &h: hour. list of integers in 0 ... 23
  • &i: interval. positive integer to apply to frequency, e.g., with @r m &i 3, repetition would occur every 3 months
  • &m: monthday. list of integers 1 ... 31
  • &M: month number. list of integers in 1 ... 12
  • &n: minute. list of integers in 0 ... 59
  • &s: set position: integer
  • &u: until. datetime
  • &w: weekday. list from SU, MO, ..., SA possibly prepended with a positive or negative integer
  • &W: week number. list of integers in 1, ..., 53



  • A link character, 'g', is appended to the flags column of any reminder with an @g goto link. Press ^g with such a reminder selected to have the operating system open the link using the default application if the link can be interpreted as a file path or a url. E.g. if the link is a URL, then it would be opened using the default browser. If link can be interpreted as a shell command followed by arguments, then the command would be executed with the arguments.

  • An in-basket character, ⓘ , is appended to the right end of the status bar when a file named 'inbasket.text' is found in the etm root directory. This file should contain lines containing etm reminder entries - one on each line. It can be imported using the import file command bound to F5.

  • An update available character, 𝕦, is appended to the right end of the status bar when checking for updates is enabled and a later version of etm is available. Details for enabling checking for updates are in configuration.

  • Alerts and beginbys can be added to any reminder with an @s start date/time entry. Alerts require a datetime in @s; beginbys also allow a date in @s.

  • A beginby is specified by adding @b n to a reminder where n is a positive integer and is interpreted as a number of days. A reminder with such an entry will be displayed in agenda view on the current date provided that the current date is no more than n days before the start date/time of the reminder. Such a warning will appear n days before, n-1 days before and so forth until 1 day before the starting date/time of the reminder. The warning displays the type character >, the summary of the reminder and the number of days remaining.

  • An alert is specified by adding @a <list of time periods>: <list of commands> to a reminder. The time periods must be given in the usual etm format, e.g., 1h13m for one hour and 13 minutes. The commands are single alphabetic characters, e.g., a, b and such. The commands used must either be e (email) or t (text) or be specified in the alerts section of the cfg.yaml file in your etm home directory. See configuration for details about this file. Basically, it associates a command, such as v with a shell command to be invoked when an alert that includes v is triggered. E.g., the alert

      	@a 20m: v

    would be triggered 20 minutes before the datetime specified in the reminder's @s entry and at that time the shell command associated with v would be invoked. Both positive (before) and negative (after) time periods can be used. Thus this entry

      	@a 1d, -1h: v, w

    would invoke the shell commands associated with v and w, one day before and again 1 hour after the datetime specified in the reminder's @s entry.

  • Reminders can have more than one @a alert entries. Different alerts could, for example, be used to trigger their commands at different times.

  • With an email, e, or text alert, t, the item summary is used as the subject and an email or text message is sent to each attendee listed in @n entries. The content of the body of the emails/messages are options that can be set in the user's configuration file.

  • Alerts and beginbys are only triggered for unfinished tasks and, when the task is repeating, only for the first unfinished instance. Similarly, pastdue notices for repeating tasks are only triggered for the first unfinished instance.


  • Daylight savings time using @s and @r. The time specified in @s is respected in the repetitions. E.g., the first five repetitions for

      * monthly @s jan 1 2020 9a @r m

    would be

          ┌─────| First 5 repetitions |──────┐
          │                                  │
          │ from Wed Jan 1 2020:             │
          │   Wed Jan 1 2020 9:00am EST      │
          │   Sat Feb 1 2020 9:00am EST      │
          │   Sun Mar 1 2020 9:00am EST      │
          │   Wed Apr 1 2020 9:00am EDT      │
          │   Fri May 1 2020 9:00am EDT      │
          │                                  │
          │           <    OK    >           │

    Note that all start at 9am even though the first 3 are EST and the last 2 are EDT.

  • Using @s with @r and with @+. Datetimes from @+ are added to the datetimes generated from the @r entry. Note that the datetime from @s will only be included if it matches one generated by the @r entry or by one included in @+. E.g.,

      * my event @s 2018-02-15 3p @r d &h 18 @+ 2018-03-02 4p

    would repeat daily at 6pm starting Feb 15 and at 4pm on Mar 2, but not at 3pm on Feb 15 which is specified in @s.

  • Using @s with @+ but without @r. Datetimes from @+ are added to @s. E.g.,

      * my event @s 2018-02-15 3p @+ 2018-03-02 4p

    would repeat at 4pm on Mar 2 and at 3pm on Feb 15 since @r is not specified.

  • Using &c and &u in @r. It is an error in dateutil to specify both &c (count) and &u (until) since providing both would be at best redundant and at worst conflicting.

    A distinction between using @c and @u is worth noting and can be illustrated with an example. Suppose an item starts at 10am on a Monday and repeats daily using either count, &c 5, or until, &u fri 10a. Both will create repetitions for 10am on each of the weekdays from Monday through Friday. The distinction arises if you later decide to delete one of the instances, say the one falling on Wednesday, using @-. With count, you would then have instances falling on Monday, Tuesday, Thursday, Friday and Saturday to satisfy the requirement for a count of five instances. With until, you would have only the four instances on Monday, Tuesday, Thursday and Friday to satisfy the requirement that the last instance falls on or before 10am Friday.

repetition examples

  • Christmas (an all day event) [r]epating (y)early on Dec 25.

      * Christmas @s 2015/12/25 @r y
  • Get a haircut (a task) on the 24th of the current month and then [r]epeatedly at (d)aily [i]ntervals of (14) days and, [o]n completion, (r)estart from the last completion date:

      - haircut @s 24 @r d &i 14 @o r
  • Take out trash (at task) on Mondays but if the task becomes [o]verdue, (s)kip the pastdue reminders.

      - Take out trash @s mon @r w @o s
  • A sales meeting (an event) [r]epeating m)onthly on [w]eekdays that are either the first or third Tuesdays in the month.

      * sales meeting @s tue 9a @e 45m @r m &w 1tu, 3tu
  • Good Friday each year 2 days before [E]aster Sunday.

      * Good Friday @s 1/1/2015 @r y @E -2
  • Friday tennis at 9:30am in November, December, January and February and at 8am in the other months:

      * Friday tennis @s 2019-01-01 6a @e 90m
        @r m &w fr &M 1, 2, 11, 12 &h 9 &n 30
        @r m &w fr &M 3, 4, 5, 6, 7, 8, 9, 10 &h 8 &n 0
  • Payday on the last week day of each month. The &s -1 part of the entry extracts the last (-1) date which is both a weekday and falls within the last three days of the month):

      * payday @s 1/1 @r m &w MO, TU, WE, TH, FR &m -1,
        -2, -3 &s -1

anniversary substitutions

Repeating events or tasks can have an anniversary expression, {XXX}, in the summary that will be replaced by the appropriate ordinal for the anniversary: 1st, 2nd, 3rd, etc. E.g.,

* {XXX} of 60 auto payments due @s 2020-06-01 @r m &c 60

would appear in agenda view

  • On July 1 2020
    • 1st of 60 auto payments due
  • On Aug 1 2020
    • 2nd of 60 auto payments due


  • On May 1 2025
    • 59th of 60 auto payments due
  • On Jun 1 2025
    • 60th of 60 auto payments due


* Will's {XXX} birthday @s 1985-08-23 @r y

would appear in agenda view on Aug 23, 2020 as

  • Will's 35th birthday

As a final example, newlyweds might want to celebrate anniversaries initially at monthly intervals

* Our {XXX} monthly anniversary @s 2020-06-18 @r m

and then eventually change to yearly celebrations by removing the 'monthly' and changing the 'm' to 'y'.

In addition to monthly and yearly frequencies, d (daily) and w (weekly) are also supported.

archived reminders

When a reminder is 'archived' in etm, it is is moved from the items table in the database to the archive table. Reminders in the archive table can only be viewed by opening query view and then beginning the query with 'a' to use the archive table. All other views display reminders from the items table.

There are two ways to archive a reminder:

  • Automatically. If 'archive_after' in the configuration settings is set to a positive integer, then tasks with finished datetimes and events with last datetimes more than this number of years before the current date will be achived automatically at the beginning of each new day. Note that unfinshed tasks, journal entries and inbox reminders are never automatically archived.
  • Manually. Select a reminder and press ^x (control and x) to archive it.

There is only one way to un-archive a reminder. Run a query beginning with 'a' in query view to use the archive table, select a reminder and press ^x. This will un-archive the reminder, i.e., move it back to the items table.


Configuration settings for etm are specified in the file cfg.yaml located in your etm home directory. See installation for the location of this directory. When etm is running, you can press F8 to open this configuration file using your system default editor for yaml (text) files. Note that any changes you make will not become effective until you stop and restart etm.

Here are the options with their default values from that file. The lines beginning with # are comments that describe the settings.

###################### IMPORTANT ########################
# Changes to this file only take effect when etm is next
# restarted.

# ampm: true or false. Use AM/PM format for datetimes if true
# else use 24 hour format.
ampm: true

# yearfirst and dayfirst. Each true or false. Whenever an
# ambiguous date is parsed, the dayfirst and yearfirst
# parameters control how the information is processed.
# Here is the precedence in each case:
#   If dayfirst is False and yearfirst is False:
#       MM-DD-YY
#       DD-MM-YY
#       YY-MM-DD
#   If dayfirst is True and yearfirst is False:
#       DD-MM-YY
#       MM-DD-YY
#       YY-MM-DD
#   If dayfirst is False and yearfirst is True:
#       YY-MM-DD
#       MM-DD-YY
#       DD-MM-YY
#   If dayfirst is True and yearfirst is True:
#       YY-MM-DD
#       DD-MM-YY
#       MM-DD-YY
yearfirst: false
dayfirst: false

# updates_interval: a non-negative integer. If positive,
# automatically check for updates every 'updates_interval'
# minutes. If zero, do not automatically check for updates.
# When enabled, a blackboard u symbol, 𝕦, will be displayed at
# the right end of status bar when an update is available
# or a question mark when the check cannot be completed
# as, for example, when there is no internet connection.
updates_interval: 0

# locale: A locale abbreviation. E.g., "en_AU" for English
# (Australia), "en_US" for English (United States), "fr_FR"
# for French (France) and so forth. Google "python locale
# abbreviatons" for a complete list."
locale: en_US

# vi_mode: true or false. Use vi keybindings for editing if
# true else use emacs style keybindings.
vi_mode: false

# secret: A string to use as the secret_key for @m masked
# entries. In etm versions after 4.0.21, the default string
# is randomly generated when this file is created or when
# the secret value is removed and etm is restarted. WARNING:
# if this key is changed, any @m entries that were made before
# the change will be unreadable after the change.
secret: %s

# omit_extent: A list of calendar names with each name
# indented on a separate line. Events with @c entries
# belonging to this list will only have their starting times
# displayed in agenda view and will neither appear nor cause
# conflicts in busy view.

# keep_current: non-negative integer. If positive, the agenda
# for that integer number of weeks starting with the current
# week will be written to "current.txt" in your etm home
# directory and updated when necessary. You could, for
# example, create a link to this file in a pCloud or DropBox
# folder and have access to your current schedule on your
# mobile device.
keep_current: 0

# keep_next: true or false. If true, the 'do next' view will
# be written to "next.txt" in your etm home directory. As with
# "current.txt", a link to this file could be created in a
# pCloud or DropBox folder for access from your mobile device.
keep_next: false

# archive_after: A non-negative integer. If zero, do not
# archive items. If positive, finished tasks and events with
# last datetimes falling more than this number of years
# before the current date will automatically be archived on a
# daily basis.  Archived items are moved from the "items"
# table in the database to the "archive" table and will no
# longer appear in normal views. Note that unfinished tasks
# and records are not archived.
archive_after: 0

# num_finished: A non-negative integer. If positive, when
# saving retain only the most recent 'num_finished'
# completions of an infinitely repeating task, i.e., repeating
# without an "&c" count or an "&u" until attribute. If zero or
# not infinitely repeating, save all completions.
num_finished: 0

# usedtime_minutes: Round used times up to the nearest
# usedtime_minutes in used time views. Possible choices are 1,
# 6, 12, 30 and 60. With 1, no rounding is done and times are
# reported as hours and minutes. Otherwise, the prescribed
# rounding is done and times are reported as floating point
# hours. Note that each "@u" timeperiod is rounded before
# aggregation.
usedtime_minutes: 1

# alerts: A dictionary with single-character, "alert" keys and
# corresponding "system command" values. Note that characters
# "t" (text message) and "e" (email) are already used.  The
# "system command" string should be a comand with any
# applicable arguments that could be run in a terminal.
# Properties of the item triggering the alert can be included
# in the command arguments using the syntax '{property}', e.g.,
# {summary} in the command string would be replaced by the
# summary of the item. Similarly {start} by the starting time,
# {when} by the time remaining until the starting time,
# {location} by the @l entry and {description} by the @d entry.
# E.g., If the event "* sales meeting @s 2019-02-12 3p"
# triggered an alert 30 minutes before the starting time the
# string "{summary} {when}" would expand to "sales meeting in
# 30 minutes". E.g. on my macbook
#    alerts:
#        v:   /usr/bin/say -v "Alex" "{summary}, {when}"
#        ...
# would make the alert 'v' use the builtin text to speech sytem
# to speak the item's summary followed by a slight pause
# (the comma) and then the time remaining until the starting
# time, e.g., "sales meeting, in 20 minutes" would be triggered
# by including "@a 20m: v" in the reminder.

# expansions: A dictionary with 'expansion name' keys and
# corresponding 'replacement string' values. E.g. with
#    expansions:
#        tennis: "@e 1h30m @a 30m: d @i personal:exercise"
#        ...
# then when "@x tennis" is entered the popup completions for
# "@x tennis" would offer replacement by the corresponding
# "@e 1h30m @a 30m: d @i personal:exercise".

# sms: Settings to send "t" (sms text message) alerts to the
# list of phone numbers from the item's @n attendee
# entries using the item's summary and the body as specified
# in the template below as the message. E.g., suppose you
# have a gmail account with email address ""
# and want to text alerts to Verizon moble phone (123)
# 456-7890. Then your sms entries should be
#     from:
#     pw:     your gmail password
#     server:
# and your item should include the following attendee entry
#     @n
# In the illustrative phone number, is the mms
# gateway for Verizon. Other common mms gateways are
#     AT&T:
#     Sprint:
#     T-Mobile:
# Note. Google "mms gateway listing" for other alternatives.
    body:   "{location} {when}"

# smtp: Settings to send "e" (email message) alerts to the
# list of email addresses from the item's @n attendee
# entries using the item's summary as the subject and body as
# the message. E.g., if you have a gmail account with email
# address "", then your entries should
# be
#     from:
#     id: whatever457
#     pw: your gmail password
#     server:
    body: "{location} {when}\\n{description}"

# queries: A dictionary with short query "keys" and
# corresponding "query" values. Each "query" must be one
# that could be entered as the command in query view. Keys
# can be any short string other than 'a', 'u', 'c' or 'l'
# which are already in use.
# unfinished tasks ordered by location
    td: m l -q equals itemtype - and ~exists f
# usedtimes by i[:1], month and i[1:2] with d
    ui: u i[:1]; MMM YYYY; i[1:2] -a d
# usedtimes by week and day for the past and current week
    uw: u WWW; ddd D -b weekbeg - 1w -e weekend
# finished|start by i[:1], month and i[1:2] with u and d
    si: s i[:1]; MMM YYYY; i[1:2] -a u, d
# items with u but missing the needed i
    mi: exists u and ~exists i
# all archived items
    arch: a exists itemtype
# items in which either the summary or the @d description
# contains a match for a RGX (to be appended when executing
# the query)
    find: includes summary d

# style: dark or light. Designed for, respectively, dark or
# light terminal backgounds. Some output may not be visible
# unless this is set correctly for your display.
style: dark

# colors: a 'namedcolor' entry for each of the following items:
#     plain:        headings such as outline branches
#     today:        the current date heading in agenda view
#     inbox:        inbox reminders
#     pastdue:      pasdue task warnings
#     begin:        begin by warnings
#     journal:      journal reminders
#     event:        event reminders
#     waiting:      waiting job reminders (unfinished prereqs)
#     finished:     finished task/job reminders
#     available:    available task/job reminders
# The default entries are suitable for the style "dark" given
# above. Note that the color names are case sensitive.
# To restore the default colors for whichever "style" you have
# set above, remove the color name for each of the items you
# want to restore and restart etm.
# To preview the namedcolors, download "" from
#    "",
# open a terminal with your chosen background color and run
#    python3 <path to>
# at the command prompt.
    plain:        'Ivory'
    today:        'Ivory bold'
    inbox:        'Yellow'
    pastdue:      'LightSalmon'
    begin:        'Gold'
    journal:      'GoldenRod'
    event:        'LimeGreen'
    waiting:      'SlateGrey'
    finished:     'DarkGrey'
    available:    'LightSkyBlue'

# locations: a dictionary with location group names and
# corresponding lists of locations. When given, do next
# view will group items first by the location group name
# and then by the location within that group. Note that
# locations can appear under more than one group name. E.g.,
# locations:
#    HOME: [home, garage, yard, phone, computer]
#    WORK: [work, phone, computer, copier, fax]
# Items with a location entru that does not belong to one
# of these location groups will be listed under 'OTHER' and
# items without a location entry under 'OTHER' and then '~'.

Note that in the 'dictionary' entries above, the components must be indented. E.g., the illustrative alert entry would be:

	v: /usr/bin/say -v "Alex" "{summary}, {when}"

data storage

All etm reminders are stored in the text file db.json in your etm home directory using the wonderful TinyDB package. This json file is human readable but not easily editable. When you start etm for the first time, this file will have no entries:

"items": {},
"archive": {}

Note that to json this is a hash/dictionary with two keys: "items" and "archive". Both have empty hashes/dictionaries as values.

Add a first reminder to take out the trash on Mondays

- trash @s 2019-12-21 @r w &w mo

at 10:26am on Dec 21, 2019 EST and the file would change to

"items": {
"1": {
"itemtype": "-",
"summary": "trash",
"s": "{D}:20191223",
"r": [
	"r": "w",
	"w": [
"created": "{T}:20191221T1526A"
"archive": {}

items now has a first item with key (unique identifier) "1". The value corresponding to this key is a hash with keys and values corresponding to the attributes of the reminder. Though these attributes are stored as strings, some actually represent non-string objects. How is this possible?

Take the entry 2019-12-23 for @s. To etm this is a pendulum date object. When etm stores this record, TinyDB recognizes that this attribute is a pendulum date object and encodes/serializes it as the string "{D}:20191223". When etm retrieves this record from storage, TinyDB recognizes from the {D} that it is to be decoded and returned as a pendulum date object. All this is completely transparent - etm gives a date object to TinyDB which stores it as a string and when etm wants it back, TinyDB deserializes it and returns it as a date object.

As another example, When the reminder is created by etm the created timestamp is handed to TinyDB as an aware pendulum datetime object with US/Eastern as the timezone. TinyDB recognizes this and, because it is an aware datetime, first converts it to Universal time and then encodes/serializes it at "{T}:20191221T1526A". The {T} indicates that it is a datetime object and the appended A indicates that it is aware and thus has been converted to Universal time. When etm retrieves this record, TinyDB recognizes from the {T} and the A that this is an aware datetime object, decodes it as an aware datetime object and, because it is aware, converts it from Universal time to whatever the current local timezone happens to be before returning it. Had this been a naive datetime, an N would have been appended to the serialization and no conversion would have been done either way.

These date and datetime serializations are extensions of TinyDB provided by etm. Three further extensions are also provided: interval for pendulum duration objects using the tag {I}, weekday for dateutil weekday objects using the tag {W} and mask using the tag {M} for encoding/serializing etm strings in a masked or obfuscated manner.

Project details

Release history Release notifications | RSS feed

This version


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for etm-dgraham, version 4.7.2
Filename, size File type Python version Upload date Hashes
Filename, size etm_dgraham-4.7.2-py2.py3-none-any.whl (167.9 kB) File type Wheel Python version py2.py3 Upload date Hashes View
Filename, size etm-dgraham-4.7.2.tar.gz (405.9 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page