Skip to main content

('A package to collect measurements published by TESS instruments into a SQlite database',)

Project description

tessdb-server(overview)

Linux service to collect measurements pubished by TESS Sky Quality Meter via MQTT. TESS stands for Cristobal Garcia's Telescope Encoder and Sky Sensor

tessdb is being used as part of the STARS4ALL Project.

Description

tessdb is a software package that collects measurements from one or several TESS instruments into database (currently a SQLite Database).

It is a Python Twisted Application that uses a custom Twisted library implementing the MQTT protocol

Desktop applicatons may query the database to generate reports and graphs using the accumulated, historic data. There are some reports scripts already included in the package, specially an IDA-format monthly report script.

Note: The Windows version has been dropped, as it was never used.

These data sources are available:

  • individual samples (real time, configurable, 1 min. aprox between samples).

Instrument should send their readings at twice the time resolution specified in the config file (in seconds).

Warning: Time referencie is always UTC, not local time.

INSTALLATION

Requirements

The following components are needed and should be installed first:

  • python 2.7.x (tested on Ubuntu Python 2.7.6) or python 3.6+

Note: It is foreseen a Python 3 migration in the future, retaining Python 2.7.x compatibility.

Installation

Installation is done from GitHub:

git clone https://github.com/astrorafael/tessdb.git
cd tessdb
sudo python setup.py install

Note: Installation from PyPi is now obsolete. Do not use the package uploaded in PyPi.

  • All executables are copied to /usr/local/bin
  • The database is located at /var/dbase/tess.db by default
  • The log file is located at /var/log/tessdb.log
  • The following required PIP packages will be automatically installed:
    • twisted,
    • twisted-mqtt
    • pyephem

Start up and Verification

  • Type sudo tessdb -k to start the service in foreground with console output. Verify that it starts without errors or exceptions. When done, abort it with ^C

From tessdb release 1.2.0, the background execution is handled as a systemd service instead of the old system V style init script:

`sudo systemctl start tessdb`

although the old sudo service tessdb start command still works.

  • It is strongly recommended to enable the service at boot time by issuing:

sudo systemctl enable tessdb

CONFIGURATION

There is a small configuration file for this service:

  • /etc/tessdb/config (Linux)

This file is self explanatory. In special, the database file name and location is specified in this file. Some of the properities marked in this file are marked as reloadable property. This means that this value can be changed and the process reloads its new value on the fly.

Sunrise / Sunset filtering

It is recommended to activate the sunrise/sunset filter to reject TESS readings coming in daytime and avoid unnecessary grouth in the database. Each location has a latitude (degrees), longitude (degrees) and elevation (meters) which can be neglected is this filter is not used.

Activating this filter have the following conecuences:

  1. Once a day, at arount 00:00 UTC, all locations will have their sunrise and sunset time computed, according to the local horizon defined (configurable).
  2. Instruments assigned to locations found to be in daytime will have their readings rejected.
  3. New: Instruments assigned to locations with NULL or Unknown longitude, latitude or elvation will have their readings accepted anyway. The only way to enable or disable writting to the database is by using the tess utility.

Logging

Log file is usually placed under /var/log/tessdb.log in Linux or under C:\tessdb\log folder on Windows. Default log level is info. It generates very litte logging at this level. File is rotated by logrotate only under Linux.

OPERATION

Server Status/Start/Stop/Restart

  • Service status: sudo systemctl status tessdb or sudo service tessdb status
  • Start Service: sudo systemctl start tessdb or sudo service tessdb start
  • Stop Service: sudo systemctl stop tessdb or sudo service tessdb stop
  • Restart Service: sudo systemctl restart tessdb or sudo service tessdb stop

Service Pause/Resume

The server can be put in pause mode, in which will be still receiving incoming MQTT messages but will be internally enqueued and not written to the database. Also, all connections to the database are closed. This allows to perform high risk operations on the database without loss of incoming data. Examples:

  • Compact the database using the SQLite VACUUM pragma
  • Migrating data from tables.
  • etc.

Service pause/resume use internally signals SIGUSR1 and SIGUSR2.

To pause the server, type: sudo tessdb_pause and watch the log file output wit tail -f /var/log/tessdb.log:

2018-11-23T13:08:25+0100 [dbase#info] TESS database writer paused
2018-11-23T13:08:25+0100 [dbase#info] Closed a DB Connection to /var/dbase/tess.db

Note: The old sudo service tessdb pause command do not work anymore.

To resume normal operation type sudo tessdb_resume and watch the same log file:

2018-11-23T13:10:37+0100 [dbase#info] TESS database writer resumed
2018-11-23T13:10:37+0100 [dbase#info] Opened a DB Connection to /var/dbase/tess.db

Service reload

During a reload the service is not stopped and re-reads the new values form the configuration file and apply the changes. In general, all aspects not related to maintaining the current connection to the MQTT broker or changing the database schema can be reloaded. The full list of reloadadble properties is described inside the configuration file.

  • Type sudo systemctl reload tessdb or sudo service tessdb reload.

Mainrteinance

Database mainteinance is made through the tess command line utility, installed by the tessdb-reports package. Mainteninance operations include:

  • create new locations
  • create new TESS instruments (manually)
  • assign locations to instruments
  • enable recording of data received from an instrument
  • listing current instruments
  • listing instruments not assigned to any known location
  • etc.

DATA MODEL

Dimensional Modelling

The data model follows the [dimensional modelling] (https://en.wikipedia.org/wiki/Dimensional_modeling) approach by Ralph Kimball. More references can also be found in Star Schemas.

The data model

The figure below shows the layout of tessdb.

TESS Database Model

Dimension Tables

They are:

  • date_t : preloaded from 2016 to 2026
  • time_t : preloaded, with seconds resolution (configurable)
  • tess_t : registered TESS instruments collecting data
  • location_t : locations where instruments are deployed
  • tess_units_t : an assorted collection of unit labels for reports, preloaded with current units.
  • tess_v : View with TESS instrument and current location. It is recommended that reporting applications use this view, instead of the underlying tess_t and location_t tables.

Date Dimension

Pretty much standard date table from dimensional modelling. Contains most used attributes plus julian_day specific to Astronomy domain.

Time of the day Dimension

Pretty much standard time of the day table from dimensional modelling. Contains well known attributes.

Instrument Dimension

This dimension holds the current list of TESS instruments.

  • The real key is an artificial key tess_id linked to the Fact table.
  • The mac_address could be the natural key if it weren't for the zero point and filter history tracking.
  • (OBSOLETE !) (To be removed soon) The name attribute could be an alternative key for the same reason. TESS instruments send readings using this name in the MQTT payload. A TESS instrument name can be changed as long as there is no other instrument with the same name.
  • The location_id is a reference to the current location assigned to the instrument. Location id -1 denotes the "Unknown" location.
  • model refers to the current TESS model.
  • firmware contains the current firmware version.
  • fov contains the Field of View, in degrees.
  • cover_offset is an additional offset in mag/arcserc^2 to account for an additional optical window attenuation attached tothe TESS itself. Defaults to 0.0.
  • channel is the current channel identifier. Default value is 0. Currently, the TESS photometer has only one channel.
  • authorised to allow the TESS instrument to store readings on the database. Authorization is a manual process done by the tess utility.
  • registered shos if the TESS instrument registered itself on the database ("Automatic") or it was done by a manual process ("Manual") using the tess utility. The default value is "Unknown" for the TESS instrument registered before adding this feature to the software. It is expected to identify these cases one by one and set them to 'Manual' or 'Automatic'.
Version-controlled Attributes

These attubutes are version-controlled and a historic of these is maintained. A new change in any of them will generate a new row in the tess_t table

  • The zero_point holds the current value of the instrument calibration constant. Defaults to '20.5' (uncalibrated photometer).
  • The filter holds the current TESS filter (i.e. 'UVIR' or Dichroic Glass). Defaults to 'UVIR'
  • The azimuth and altitude attributes hold the photometer current orientation, in degrees. Default azimuth is 0.0 and default altitude is 90.0
Version Control Attributes

These columns manage the version control of a given TESS attributes.

  • Columns valid_since and valid_until hold the timestamps where the changes to version controlled attributes are valid.
  • Column valid_state is an indicator. Its values are either Current or Expired. The current valid TESS instrument has its valid_state set to Current and the expiration date in a far away future (Y2999).

Unit dimension

The tess_units_t table is what Dr. Kimball denotes as a junk dimension. It collects various labels denoting the current measurement units of samples in the fact table.

  • Columns valid_since, valid_until and valid_state keep track of any units change in a similar technique as above should the units change.

Location dimension

This dimension table holds all known locations where TESS photometers are to be deployed.

  • site. Unique site name describing the this location.
  • contact_person. Person to account for observations.
  • organization. Organization where the contact person belongs to or running the facilities in the location.
  • contact_email. Contact person email address.
  • longitude Location longitude in degrees. West is negative.
  • latitude. Location latitude in degrees
  • elevation. Location elevation in meters
  • zipcode. Location ZIP code
  • province. Location country
  • country. Location country
  • timezone. Time zone (to calculate local time) in standard format described by Wikipedia[https://en.wikipedia.org/wiki/List_of_tz_database_time_zones]
  • sunrise & sunset. Computed attributes (oce per day) used to filter out readings in daylight.

Fact Tables

They are:

  • tess_readings_t : Accumulating snapshot fact table containing measurements from several TESS instruments.

TESS devices with accelerometer will send azimuth and altitude values, otherwise they are NULL.

TESS devices with a GPS will send longitude, latitude and height values, otherwise they are NULL.

Other Tables

It is possible now to replace a given TESS-W and keep the name. This is sueful for sites that wish to announce a simbolic name for its TESS-W photometer and never change even if the device is replaced (i.e. by being broken).

So now, a given name does not identify a TESS-W photometer, only the MAC address does. An association table is needed

  • name_to_mac_t : Association table to label a given TESS-W device (identified by its MAC address) to a symbolic name.

Sample SQL Queries

The following are samples queries illustraing how to use the data model. They are actually being used by the STARS4ALL project

  1. Get a daily report of readings per instrument:
#!/bin/bash
sqlite3 /var/dbase/tess.db <<EOF
.mode column
.headers on
SELECT d.sql_date, i.name, count(*) AS readings
FROM tess_readings_t AS r
JOIN tess_t AS i USING (tess_id)
JOIN date_t AS d USING (date_id)
GROUP BY r.date_id, r.tess_id
ORDER BY d.sql_date DESC;
EOF
  1. Extract a CSV (semicolon separated) with all readings for an instrument passed as a command line argument:
#!/bin/bash
instrument_name=$1
sqlite3 -csv -header /var/dbase/tess.db <<EOF
SELECT (d.julian_day + t.day_fraction) AS julian_day, (d.sql_date || 'T' || t.time) AS timestamp, r.sequence_number, l.site, i.name, r.frequency, r.magnitude, i.zero_point, r.sky_temperature, r.ambient_temperature
FROM tess_readings_t AS r
JOIN tess_t          AS i USING (tess_id)
JOIN location_t      AS l USING (location_id)
JOIN date_t          AS d USING (date_id)
JOIN time_t          AS t USING (time_id)
WHERE i.name = "${instrument_name}"
ORDER BY r.date_id ASC, r.time_id ASC;
EOF
  1. Show current TESS instruments. Note that we are using the tess_v View,so that the current location info is already included.
#!/bin/bash
sqlite3 /var/dbase/tess.db <<EOF
.mode column
.headers on
SELECT v.name AS Name, v.mac_address AS MAC, (v.latitude || ' ' || v.longitude) AS Coordinates , (v.site || ', ' || v.location || ', ' || v.province) AS Location, v.contact_email as User, v.zero_point as ZP, v.filter as Filter
FROM tess_v AS v
WHERE v.valid_state = "Current"
ORDER BY v.name ASC;
EOF
  1. Show TESS instruments changes (zero point and/or filter)
#!/bin/bash
sqlite3 /var/dbase/tess.db <<EOF
.mode column
.headers on;
SELECT i.name AS Name, i.zero_point as ZP, i.filter as Filter, i.valid_since AS Since, i.valid_until AS Until, i.valid_state AS 'Change State'
FROM tess_t AS i
ORDER BY i.name ASC, i.valid_since ASC;
EOF
  1. Show the time span of readings per TESS
#!/bin/bash
sqlite3 /var/dbase/tess.db <<EOF
.mode column
.headers on;
SELECT i.name, MIN(d.sql_date || 'T' || t.time || 'Z') AS earliest, MAX(d.sql_date || 'T' || t.time || 'Z') AS latest
FROM tess_readings_t AS r
JOIN tess_t          AS i USING (tess_id)
JOIN location_t      AS l USING (location_id)
JOIN date_t          AS d USING (date_id)
JOIN time_t          AS t USING (time_id)
GROUP BY i.name;
EOF
  1. Show locations not assigned to any TESS
#!/bin/bash
sqlite3 /var/dbase/tess.db <<EOF
.mode column
.headers on;
SELECT l.site
FROM location_t        AS l 
LEFT OUTER JOIN tess_t AS i USING (location_id)
WHERE i.name IS NULL;
EOF

MQTT PAYLOAD INFORMATION

Payloads are transmitted in JSON format, with the format described below.

Published on topic 'STARS4ALL/{channel}/reading'

Field name Type Units Optional Description
seq int - mand Sequence number. If possible use 32 bits. The sequence number will start in 1 at each device reboot.
name string - mand Instrument friendly name. Should be unique as it identifies the device.
freq float Hz mand Raw reading as a frequency with 3 decimal digits precision (millihertz) NNNNN.NNN
mag float mag/arcsec^2 mandat. Visual magnitude (formulae?) corresponding to the raw reading). Transmitted up to two decimal places NN.NN
tamb float ºC mandat. Ambient Temperature. Transmitted up to one decimal place.
tsky float ºC mandat. Sky Temperature. Transmitted up to one decimal place.
wdBm int dBm opt WiFi Received Signal Strength.
az int deg opt Photometer optical axis Azimuth sent only on instruments with accelerometer.
alt int deg opt Photometer optical axis Altitude (angle): sent only on instruments with accelerometer.
lat float deg opt Instrument latitude. Only sent by instruments with GPS integration.
long float deg opt Instrument longitude. Only sent by instruments with GPS integration.
height float meters opt Instrument height above the sea level. Only sent by instruments with GPS integration.
rev int - mand Payload data format revision number. Current version is 1.
tstamp string UTC opt Timestamp,“YYYY-MM-DDTHH:MM:SS” format. Must be UTC.

Published on topic 'STARS4ALL/register'

Field name Type Units Optional Description
name string - mand Instrument friendly name. Should be unique as it identifies the device.
mac string - mand. Device MAC address, format “xx:yy:zz:rr:ss:tt”
calib float mag/arcsec^2 mand Per-device Zero Point. Transmitted as NN.NN floating point.
rev int - mand Payload data format revision number. Current version is 1.
chan string - opt Channel where this instrument will publish its readings.

OPERATION & MAINTENANCE

Pause & resume

Since the current database used is SQLite - a single user database - you need to pause tessdb-server if dealing directly with the database like issuing SQL commands directly or using the tess command line utility

Use /usr/local/bin/tessdb_pause and /usr/local/bin/tessdb_resume to coordinate your direct interactions to the database with tessdb-server.

Reload

Since tessdb-server maintains a RAM cache of photometers data, some tess command requires a server reload to rfefress the cache. The tessutility wanrs you when this is necessary.

Restart

The newest filter operation mode in tessdb-server maintains a sliding window of photometers samples before writting to database If it is necessary to restart the server, use /usr/local/bin/tessdb_restart instead of service tessdb restart. This will ensure that the lastest readings are stored in the database.

The tess utility

tess is a command line utility to perform some common operations on the database without having to write SQL statements. As this utility modifies the database, it is necessary to invoke it within using sudo. Also, you should ensure that the database is not being written by tessdb to avoid database is locked exceptions, either by using it at daytime or by pausing the tessdb service with /usr/local/bin/tessdb_pause and then resume it with /usr/local/bin/tessdb_resume.

Project details


Release history Release notifications

Download files

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

Files for tessdb-server, version 2.6.0
Filename, size File type Python version Upload date Hashes
Filename, size tessdb-server-2.6.0.tar.gz (63.5 kB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page