Preprocessor for scraped llvm/llvm-project PRs & comments → documents for Pinecone

[jonathanMLDev]

Summary

Preprocess already scraped llvm/llvm-project pull request data (including comments) into document objects (content + metadata) for Pinecone upsert. No scraping or GitHub API calls.

Scope

• Input: Scraped PR data for llvm/llvm-project (with comments). Output: Documents with content and metadata (e.g. repo, number, state, author, created_at, labels, url). In scope: parse/validate, normalize text, include comments in content/chunking, document schema. Out of scope: GitHub fetch; Pinecone API.

Result

Library or CLI: scraped payload(s) → list of { content, metadata }. Config for field mapping/truncation. Code, tests, and doc schema README.

Acceptance criteria

• Accepts scraped llvm/llvm-project PR (+ comments) data in agreed format; outputs stable content + metadata.
• Metadata includes identifier, repo (llvm/llvm-project), type (PR), filterable fields. Schema documented.
• No GitHub API or scraping in this component.

[jonathanMLDev]

#93