Skip to content

fixed 1 big chuk issue by changing approach - #10

Open
gguillermomendoza wants to merge 27 commits into
camunity:mainfrom
gguillermomendoza:main
Open

fixed 1 big chuk issue by changing approach#10
gguillermomendoza wants to merge 27 commits into
camunity:mainfrom
gguillermomendoza:main

Conversation

@gguillermomendoza

Copy link
Copy Markdown

Fixes issue with resume chunking by locating bullet markers to group text correctly, stripping markdown headers and bold text via regex, fallback to per-line chunking when no bullets exist, added warning for low chunk count (< 5)
Changes:

  • Modified chunking logic in Step 1
  • Added test case for chunking parity
    Tests:
  • _test_chunking_parity: verifies identical chunk counts and text for plain text and markdown resumes (verified passing)

gguillermomendoza and others added 27 commits July 16, 2026 15:53
…d by locating bullet markers in the text and grouping content up to the next marker into one chunk. md headers and bold are stipped before length check. added a test and a warning is printed when a degenerate chunk set is detected
Streamlit's default styling read as generic AI-generated app chrome; Gradio
gives a cleaner, more configurable surface for the same pipeline.
removed artifact from adding with glob
added resolve_writing_sample, to take in uploads, added deterministic tests covering pasted text, .txt and .md uploads, file precedence, empty inputs, mocked PDF extraction, and unsupported uploads
…_builder

add writing sample resolution function

- added f(x)s to support uploads with uploaded files takig precedence over
pasted text and timmed empty-safe pasted-text fallback.
- resume and writing sample pdfs both delegate to
  pipeline.extract_text_from_pdf, txt files are read as UTF-8
- added determinitstic offline tests covering pasted text, file uploads,
  folse precedence, empty inputs, mocked pdf extraction, and unsupported
  uploads
* Implemented _build_cover_letter_prompt with default concise length
* Applied word, paragraph, and experience-count targets
* JSON-encoded and delimited requirements and user data
* Added strict formatting instructions and public generate_cover_letter function
* Included deterministic offline tests for data handling and generation
…ort for various claim checks, along with unit tests and Markdown report instructions
…ipeline-orchestrator

Add cover-letter pipeline orchestrator
…ab-to-gradio_app

Add cover letter workflow and tab to Gradio app
…r-letter-features

Strengthen cover-letter test coverage assertions
…nd-create-.env.example

Update and standardize documentation
…ects-in-pipeline-files

Fix cover-letter prompt and run isolation defects
…tor-to-gradio_app.py

Unify résumé and cover-letter workflow inputs. Tests Pass.
…-inputs-in-gradio_app.py

Refine UI labels/visibility for job/resume inputs and add config-based UI test
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant