Add Gemini provider, custom URLs, local providers, and extra headers

[dpage]

Summary

• Shared provider infrastructure: Extract duplicated code from OpenAI/Voyage into provider_common.c (curl, JSON, API key loading, response parsing). Slims existing providers from ~516 lines each to ~163-187 lines.
• Google Gemini provider: New provider_gemini.c with x-goog-api-key auth, model-in-URL-path, and native batch support via batchEmbedContents.
• Custom base URLs: api_url GUC now defaults to empty; each provider supplies its own default URL. Enables pointing any provider at a custom endpoint.
• OpenAI-compatible local providers: API key is optional when a custom URL is set, enabling LM Studio, Docker Model Runner, llama.cpp, and EXO.
• Extra HTTP headers: New extra_headers GUC (semicolon-separated key: value pairs) for proxy servers like Portkey.

Test plan

• make clean && make builds cleanly with zero warnings
• make install && make installcheck passes all existing tests plus new provider/GUC test cases
• Verify Gemini embeddings work with a real API key
• Verify OpenAI-compatible local provider (e.g., LM Studio) works without an API key
• Verify extra headers are sent correctly (e.g., via Portkey or request inspection)

🤖 Generated with Claude Code

[codacy-production]

Up to standards ✅

🟢 Issues 2 high · 12 medium

Results:
14 new issues

Category	Results
ErrorProne	2 high (2 false positives)
Complexity	12 medium

View in Codacy

🟢 Metrics 110 complexity · -16 duplication

Metric	Results
Complexity	110
Duplication	-16

View in Codacy

_{TIP This summary will be updated as you push new changes. Give us feedback}

[dpage]

Note that the macOS CI failure is a separate, unrelated issue.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 247624e2-df0b-43b3-8cce-c844d34d12fb

📥 Commits

Reviewing files that changed from the base of the PR and between 125396c and 3f3d5e8.

📒 Files selected for processing (1)

• TODO.md

💤 Files with no reviewable changes (1)

• TODO.md

---

📝 Walkthrough

Walkthrough

Adds a new Gemini embedding provider and registers it in the provider registry. Introduces shared provider utilities via src/provider_common.c and src/provider_common.h, and refactors OpenAI, Voyage, and Ollama providers to use those helpers. Adds pgedge_vectorizer.extra_headers GUC and changes pgedge_vectorizer.api_url default to empty. Declares pgedge_vectorizer_extra_headers and GeminiProvider in headers, updates Makefile build objects, expands tests (test/sql/providers.sql), updates docs for providers and configuration, and removes the LLM section from TODO.md.

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly summarizes the main changes: addition of Gemini provider, custom URL support, local provider compatibility, and extra headers functionality.
Description check	✅ Passed	The description is well-structured and directly related to the changeset, detailing shared provider infrastructure refactoring, the new Gemini provider, custom base URLs, local provider support, and extra headers—all of which are reflected in the code changes.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feature/llm-provider-enhancements

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}