Skip to content

feat: Add Indeed.com crawling and improve crawling prompts - #66

Merged
KevinJBoyer merged 3 commits into
kb/crawl-job-3-4from
kb/update-prompts
Oct 29, 2025
Merged

feat: Add Indeed.com crawling and improve crawling prompts#66
KevinJBoyer merged 3 commits into
kb/crawl-job-3-4from
kb/update-prompts

Conversation

@KevinJBoyer

@KevinJBoyer KevinJBoyer commented Oct 29, 2025

Copy link
Copy Markdown

Ticket

https://navalabs.atlassian.net/browse/DST-1253

Changes

  • Move "two sentence" specifier from description from Pydantic model to the extract_supports prompt (and update the prompt version)
  • Add a prompt for crawl_indeed
  • Update the prompt for crawl_gcta to provide a more informative description
  • Add a helper utility to the makefile to easily update and add crawl jobs

Context for reviewers

See ticket

Testing

make copy-prompts
make add-crawl-jobs
make process-crawl-jobs
image Screenshot 2025-10-29 at 8 50 34 AM

Preview environment for frontend

♻️ Environment destroyed ♻️

Preview environment for app

♻️ Environment destroyed ♻️

Copilot AI review requested due to automatic review settings October 29, 2025 12:56
@KevinJBoyer
KevinJBoyer changed the base branch from main to kb/crawl-job-3-4 October 29, 2025 12:56
emails: list[str]
addresses: list[str]
phone_numbers: list[str]
description: str | None = Field(description="2-sentence summary, including offerings")

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Moved into the prompt for extract_supports in Phoenix

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull Request Overview

This PR updates prompt versions in the application configuration and removes field-level documentation from the SupportEntry model. The changes appear to be coordinating updates to LLM prompt versions with a simplified schema definition.

  • Updated prompt version IDs for extract_supports and crawl_gcta operations
  • Added new prompt version for crawl_indeed operation
  • Removed the Field description metadata from the description field in SupportEntry

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File Description
app/src/ingestion/support_entry.py Removed Field description metadata for the description attribute in the SupportEntry model
app/src/app_config.py Updated prompt version IDs for existing operations and added new crawl_indeed prompt configuration

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

addresses: list[str]
phone_numbers: list[str]
description: str | None = Field(description="2-sentence summary, including offerings")
description: str | None

Copilot AI Oct 29, 2025

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The removal of the Field description metadata ('2-sentence summary, including offerings') eliminates important guidance for LLM-based extraction. The SUPPORT_ENTRY_SCHEMA on line 15 uses model_json_schema() which includes field descriptions in the schema. This description helps the LLM understand what type of content to generate for this field. Consider restoring the Field description or documenting why this guidance was removed, especially since the prompt versions are being updated in this same PR.

Suggested change
description: str | None
description: str | None = Field(description="2-sentence summary, including offerings")

Copilot uses AI. Check for mistakes.

@yoomlam yoomlam left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Missing Makefile? make add-crawl-jobs doesn't work

@yoomlam yoomlam left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nice!

@KevinJBoyer
KevinJBoyer merged commit 3d2324f into kb/crawl-job-3-4 Oct 29, 2025
10 of 11 checks passed
@KevinJBoyer
KevinJBoyer deleted the kb/update-prompts branch October 29, 2025 14:50
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.

4 participants