Wayback Tweets v0.4.x Released
Stargazers and some parses
Edited July 28, 2024
Tool that displays multiple archived tweets on Wayback Machine, to avoid opening each link manually via Wayback CDX Server API. Try it here.
Stars and Shares
In November, I shared the tool on the Bellingcat Discord server, and to my surprise, they shared it across all social media platforms, directly contributing to an increase in stars on the GitHub repository (something crucial for an independent developer like me).
In a conversation with data scientist Tristan Lee, who provided the idea for Wayback Tweets (WBT), we discussed filters and some necessary parses that I should implement later. However, the most valuable feedback came from Nathan Patin, who pointed out some inconsistencies in the URLs of old tweets, which have been addressed in the current version.
After all this buzz on social media, I decided to create a web app using the ideal technologies (React library) to make WBT more user-friendly. The project planning, which involves mapping the main UIs and URL patterns of Twitter since 2006, is currently in progress.
Additionally, the tool, currently hosted on Streamlit Cloud, received an additional 7 GiB from the Streamlit team (thanks to Jessica Smith) and is expected to be featured soon in the framework's website gallery.
What's Changed (v0.4.x)
-
0.4.3
-
Add 8-digit collapsing strategy (one capture per day)
-
Add messages about collapsing strategy and number of tweets displayed
-
0.4.2
-
Parse tweet URLs to delete
/photos
,/likes
,/retweets
and other sub-endpoints- Only for "original url"
-
0.4.1
-
Add a warning message for non 200/300 status code
-
Set a fixed tweets per page (25) due the API rate limit
-
0.4
-
Parse old tweets URLs
- Picture:
twimg.com
- Reply
username/status/"/user_reply/status/user_reply_msg_ID"
- Picture:
- Change filter text "Only deleted tweets" to "Original URLs not available" with a help info
- Change "tweet" text to "original link" on each header
-
Allows MIME type
warc/revisit
andunk
(to be reviewed) -
v0.3 released in Nov 14
-
Added:
- Filter by year
- Filter by range size
- Spinner to load data
f-string
to code
-
Updated:
- Streamlit version to 1.27.0
- Style (font, BG color)
- README
- Fix MIME type display logic
- Fix pagination
- Fix error messages
- Fix JSON response
-
Deleted:
- Progress bar
Pull Requests
- Collapse by timestamp #17
- Delete sub-endpoints for "original url" #16
- Update tweets page #15
- Fix old tweets URLs and minor fixes #14
- v0.3 #8