nothing-less 0.3.0

A TUI paging application with enhanced support for tabular data and real-time streaming

Nothing-less (nless)

Nless is a TUI paging application (based on the awesome Textual library) with vi-like keybindings. Nless has enhanced functionality for parsing tabular data:

• inferring file delimiters
• delimiter swapping on the fly
• regex-based parsing of raw logs into tabular data using Python's regex engine
• filtering
• sorting
• searching
• real-time event parsing.

Why?

As a kubernetes engineer, I frequently need to interact with streaming tabular data. k get pods -w, k get events -w, etc. I want a TUI tool to quickly dissect and analyze this data - and none of the existing alternatives had exactly what I wanted:

• streaming support
• delimiter inference - I don't want to do a bunch of work to tell the program what type of data it's viewing, I want it to infer it if possible
• vi-like keybindings So I decided to build my own tool, integrating some of my favorite features that I've seen in other similar tools.

Goals

This project is not meant to be a replacement/competitor for any of the tools mentioned in the alternatives section at the end. Instead, it's meant to bring its own unique set of features to compliment your workflow.

• UX:
  • vi-like keybindings, familiar to any VIM user
  • minimize the number of keypresses to analyze a dataset
• Kubernetes support:
  • support for K8s usecases out of the box - such as parsing data streams from kubectl
• Tabular data toolkit:
  • broad support for a variety of use-cases analyzing,filtering,sorting, and searching tabular data
  • converting data streams into tabular data, such as JSON log parsing

Getting started

Dependencies

• python>=3.13

Installation

pip install nothing-less

Usage

• pipe the output of a command to nless to parse the output $COMMAND | nless
• read a file with nless nless $FILE_NAME
• redirect a file into nless nless < $FILE_NAME
• Once output is loaded, press ? to view the keybindings

Demos

Basic functionality

The below demo shows basic functionality:

• starting with a search /
• applying that search &
• filtering the selected column by the value within the selected cell F
• swapping the delimiter D (raw and ,)

Streaming functionality

The below demo showcases some of nless's features for handling streaming input, and interacting with unknown delimitation:

• The nless view stays up-to-date as new log lines arrive on stdin (allows pipeline commands, or redirecting a file into nless)
• Showcases using a custom (Python engine) regex, example - {(?P<severity>.*)}\((?P<user>.*)\) - (?P<message>.*) - to parse raw logs into tabular fields.
• Sorts, filters, and searches on those fields.
• Flips the delimiter back to raw, sorts, searches, and filters on the raw logs

Features & Functionality

Buffers:

• All mutating actions will apply the action by replicating the current "buffer". This allows you to jump up and down the stack to see how you've analyzed your data.
• [1-9] - will select the buffer at the index corresponding to the input number
• L - selects the next buffer
• H - select the previous buffer
• q - closes the current active buffer, or the program if all buffers are closed
• N - creates a new buffer from the original data

Navigation:

• h - move cursor left
• l - move cursor right
• j - move cursor down
• k - move cursor up
• 0 - jump to first column
• $ - jump to final column
• g - jump to first row
• G - jump to final row
• w - move cursor right
• b - move cursor left
• ctrl+u - page up
• ctrl+d - page down
• c - to select a column to jump the cursor to

Column visibility

• C - will prompt for a regex filter to selectively display columns, or all to see all columns. TIP: use a non-existing column (none, for example) to only see the current pivots/count
• > - will move the current column one to the right
• < - will move the current column one to the left

Pivoting

• U - will mark the selected column as part of a composite key to group records by, adding a count column pinned to the left
  • enter - pressing enter while the cursor is over one of the composite key columns will "dive in" to the data set behind the pivot - applying the composite key as a filter in a new buffer

Filtering:

• f - will filter the current column and prompt for a filter
• F - will filter the current column by the highlighted cell
• | - will filter ALL columns and prompt for a filter
• & - applies the current search as a filter across all columns

Searching:

• / - will prompt for a search value and jump to the first match
• * - will search all columns for the current highglighted cell value
• n - jump to the next match
• p - jump to previous match

Output:

• W - will prompt for a file to write the current buffer to. - can be used to write to stdout, allowing you to use nless inside of a command chain cat $MY_FILE.txt | nless | grep -i active for example.

Sorting:

• s - toggles ascending/descending sort on the current column

json:

• in addition to the json delimiter that can be set per session or per column, there's also support for json actions:
• J - will prompt you to select a json field, under the current cell, to add as a column for further filtering/sorting/etc

Delimiter/file parsing:

• By default, nless will attempt to infer a file delimiter from the first few rows sent through stdin. It uses common delimiters to start - ,, , |, \t, etc.

• D - you can use D to explicitly swap the delimiter on the fly. Just type in one of the common delimiters above, and the rows will be re-parsed into a tabular format.

• D - alternatively, you can pass in a regex with named capture groups. Those named groups will become the tabular columns, and each row will be parsed and split across those groups. Example {(?P<severity>.*)}\((?P<user>.*)\) - (?P<message>.*)

• D - additionally you can just pass the word raw to see the raw lines behind the data. You can still sort, filter, and sarch the raw lines.

• D - pass the word json to parse the first set of keys from each JSON line (or read the whole buffer in as a JSON object/list)

• D - last, you can pass a delimiter value of (two spaces). This will parse text that has been delimited utilizing multiple spaces, while preserving values that have a single space. This is most commonly useful for parsing kubernetes output (kubectl get pods -w), for example.

• d - transforms a column into more columns using a columnar delimiter (currently json is the only delimiter supported)

Contributing

Contributions are welcome! Please open an issue or a pull request - check out the contributing guidelines for more information.

Alternatives

Shout-outs to all of the below wonderful tools! If my tool doesn't have what you need, they likely will:

• visidata
• csvlens
• lnav
• toolong