Skip to main content

datasette powered notebook search

Project description

nbsearch

Datasette based notebook search extension, originally inspired by Simon Willison's Fast Autocomplete Search for Your Website.

Binder

Related: Sketching a datasette powered Jupyter Notebook Search Engine: nbsearch

Installation

pip3 install --upgrade git+https://github.com/ouseful-testing/nbsearch.git

Usage - inside Jupyter notebook environment

This is still very much a work in progress. I think it works in Binder (erm, maybe...)

Installing the extension and restarting the Jupyter server should make a Jupyter proxy served version of the search form available from the New menu.

At the moment, I think you need to create the index explicitly. From the command line:

nbsearch index -p PATH

(From a notebook command cell, prefix with !. Default path is .).

The sqlite db that stores the results is in ~/.nbsearch/notebooks.sqlite.

All notebooks down from PATH will be indexed.

When you have generated an index, launch the nbsearch panel from the notebook homepage New menu.

Usage - from the command line

  • create a database by passing a path to some notebook files, eg:
    • nbsearch index -p "/Users/myuser/Documents/content/notebooks"
  • run the server, eg:
    • nbsearch serve

datasette should start up and display a server port number. (To kill it, I look for process IDs: ps -al |grep datasette; there is probably a better way... It might be nice if CLI kept track of process IDs and let you kill from a selection?)

UI

Old screenshot

A copy button on a code cell lets you copy code from the code cell.

Results are limited in length; the Show all cell button expands the result cell to its full length. The Show previous cell and Show next cell buttons display the full previous / next cell (repeatedly clicking these buttons grab the next next and previous previous cells etc.)

Clicking on the notebook structure visualisation graphic (the pink/blue image: the colours representent cell type and relative length) will collapse / reveal the display of the result block.

Known Issues

The links to notebooks may well be broken in the search results: I need to think about how to index and handle paths in links, particular in proxy server case.

The app requires the latest version of datasette from the repo, not pypi.

The index is not updated ever unless you rerun the indexer, although I've started trying to ponder a filesystem watchdog here. Another possibility my be a Jupyter notebook content manager or post-save hook to update records as the notebook server saves them but this would not catch filesystem operations (dragging new notebooks to a folder, deleting notebooks etc?)

If the sqlite db is updated, I assume by some magic that the datasette server queries over the updated content?

Useful

Quick way to kill datasette processes: ps aux | grep datasette | grep -v grep | awk '{print $2}' | xargs kill

Project details


Download files

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

Source Distribution

nbsearch-0.0.4.tar.gz (9.1 kB view details)

Uploaded Source

Built Distribution

nbsearch-0.0.4-py3-none-any.whl (43.0 kB view details)

Uploaded Python 3

File details

Details for the file nbsearch-0.0.4.tar.gz.

File metadata

  • Download URL: nbsearch-0.0.4.tar.gz
  • Upload date:
  • Size: 9.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.2.0 pkginfo/1.6.1 requests/2.24.0 setuptools/50.3.2 requests-toolbelt/0.9.1 tqdm/4.51.0 CPython/3.8.6

File hashes

Hashes for nbsearch-0.0.4.tar.gz
Algorithm Hash digest
SHA256 ef480823c919642d83951b3f5490143210c50583bc54c2d6a49dcb56b9693bf8
MD5 476c3d87ae3dac2d4a8b9b3ce2639e8f
BLAKE2b-256 42168beefae93970a2dc0cbae00509c634c0cc339b6aef7b03f13f72f79d74e2

See more details on using hashes here.

File details

Details for the file nbsearch-0.0.4-py3-none-any.whl.

File metadata

  • Download URL: nbsearch-0.0.4-py3-none-any.whl
  • Upload date:
  • Size: 43.0 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.2.0 pkginfo/1.6.1 requests/2.24.0 setuptools/50.3.2 requests-toolbelt/0.9.1 tqdm/4.51.0 CPython/3.8.6

File hashes

Hashes for nbsearch-0.0.4-py3-none-any.whl
Algorithm Hash digest
SHA256 c88cbcbacae3036d3368733436419d3eb41575bb4c3234f3d4ca62f65850150f
MD5 87e634d60926a7d69851090df5bfd115
BLAKE2b-256 c78d440fbd1141927b9b8ea4c9accb21fbe77199d51648d57b63bdb699ef863b

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page