(bug): S3 Cache Key Ignores max_file_size Parameter, Leading to Stale Results #531
Description
Which interface did you use?
Web UI
Repository URL (if public)
https://gitingest.com/man-group/dtale/tree/master/dtale/dash_application
Git host
GitHub (github.com)
Other Git host
No response
Repository visibility
public
Commit, branch, or tag
default branch
Did you ingest the full repository or a subdirectory?
subdirectory
Operating system
Windows
Browser (Web UI only)
Chrome
Other browser
No response
Gitingest version
No response
Python version
No response
Bug description
The S3 caching mechanism does not include the max_file_size parameter in its cache key generation.
This causes a significant issue: if a user performs an ingest with one file size limit, and then performs a second ingest on the same repository with a different file size limit, the system incorrectly returns the cached result from the first request.
The user sees stale data and is unable to get an updated digest reflecting their new settings, which can be very confusing. The root cause appears to be in server/s3_utils.py, where the hash for the cache key (s3_file_path) is generated without considering the file size limit.
Steps to reproduce
Navigate to the Gitingest Web UI using this URL: https://gitingest.com/man-group/dtale/tree/master/dtale/dash_application
Observe that the "Include files under" slider is at its default position, corresponding to a low value like 50kB.
Click the "Ingest" button and wait for the results to load.
In the "Directory Structure" output, notice that the file dtale/dash_application/layout/layout.py (which is 104kB) is correctly excluded because it is larger than the 50kB limit.
Now, move the "Include files under" slider all the way to the right, to a high value (e.g., 1MB).
Click the "Ingest" button again.Expected behavior
After the second ingest with the increased file size limit, the system should re-process the repository. The new "Directory Structure" should now include the file dtale/dash_application/layout/layout.py, as its size (104kB) is well within the new 1MB limit.
Actual behavior
The results from the second ingest are identical to the first. The "Directory Structure" still excludes dtale/dash_application/layout/layout.py, incorrectly ignoring the user's updated file size setting. The application returns a stale, cached result instead of processing the request with the new parameters.
Additional context, logs, or screenshots
This bug can be fixed by updating the generate_s3_file_path function in src/server/s3_utils.py to include the max_file_size (in KB) as part of the string that is hashed to generate the cache key.y