svandragt / fafi Goto Github PK

https://www.alfredforum.com/topic/15914-script-filter-help-searching-simplenote-via-cli/

Links containing errors should be skipped for a number of hours as they slow down the indexing.

https://developers.google.com/safe-browsing/v4/update-api

Avoid pulling in malware and help users discover malicious links.

Must preserve privacy of the user.

Dependabot couldn't authenticate with https://pypi.python.org/simple/.

You can provide authentication details in your Dependabot dashboard by clicking into the account menu (in the top right) and selecting 'Config variables'.

View the update logs.

default to application data (using appdirs).
We can always add a cli option to set our own.

I deleted my firefox profile and fafi is trying to use a profile that no longer exists.
Handle profile deletion by asking the user to pick a new sqlite.
This should not affect already indexed bookmarks.

Spike:

1. Screenshot shows previous name for the program.

Currently the default is 7 results, but it should be customisable through a commandline parameter.

404s should be skipped.
5xx should be errored
Not sure about following redirects.

blocks further dev

if we use direnv to set the vaiable to the poetry environment, then have the spec file read the env var, then the spec file can be included into version control

blocked by #9

Firefox supports multiple profiles. Therefore fafi should be able to index bookmarks relating to any of the available profiles (no parameter), or a specific profile when one is specified.

spin up a basic web server and a search/results UI

task runner, need to first publish my taskfile fork.

1. delete existing duplicate urls
2. prevent new one being saved

would be nice to install fafi using pipx install fafi or something like it which would abstract away all the environmental hoops.

see https://pypi.org/project/pipx/

• Update go
• https://goreleaser.com/quick-start/

• how to build and submit pypi updates using poetry: poetry build && poetry publish
• pyinstaller works also but is not documented:

pyinstaller --onedir src/fafi/main.py --name fafi
cd dist
mv fafi app
ln -s app/fafi fafi
zip -r ../fafi.zip
mv ../fafi.zip .

Should be created in central location (already have appdirs).
This is an issue when installing fafi via pipx.

INdexing is currently done one link at a time, we can speed it up using multiple indexers.

https://github.com/marketplace/actions/cache

1. Listview with results
2. Tab for Search and Index
3. For each result: button to open the website
4. For each result: button to read the text only scraped content

• --list

need to keep the url to avoid reimporting, but the title, text can be cleared.
might require a new field (isDeleted) to avoid indexing.

https://stackoverflow.com/questions/28895673/subcommand-alternative-to-argparse-and-optparse

search and index.

Index bookmarks old to new, and save the last indexed bookmark created date as state.
When indexing, index only bookmarks newer than that date.