docs: fixed embeddings and sync config#180
Conversation
📝 WalkthroughWalkthroughDocumentation updates systematize AI provider configuration in GitHub Actions workflows, remove browser rendering attributes from LikeC4 embeds, clarify token usage for source PRs versus model repositories across integrations, document a new LikeC4 formatting configuration variable, and restructure publish action documentation. ChangesDocumentation Configuration and Integration Updates
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 inconclusive)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Tip 💬 Introducing Slack Agent: The best way for teams to turn conversations into code.Slack Agent is built on CodeRabbit's deep understanding of your code, so your team can collaborate across the entire SDLC without losing context.
Built for teams:
One agent for your entire SDLC. Right inside Slack. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@packages/web/src/content/docs/docs/integrations/github-actions.md`:
- Line 76: The table entry for the `ai-provider` option is inconsistent with the
README and workflow examples; change the default value shown from `anthropic` to
`gemini` so `ai-provider`'s default matches the README and the updated
workflows, and scan other docs mentioning `ai-provider` to ensure all defaults
are consistently `gemini`.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 2a326178-ad20-4b72-84e2-b795b89a792b
📒 Files selected for processing (10)
README.mdpackages/web/src/content/docs/docs/contributing/index.mdpackages/web/src/content/docs/docs/how-it-works.mdpackages/web/src/content/docs/docs/integrations/github-actions.mdpackages/web/src/content/docs/docs/integrations/index.mdpackages/web/src/content/docs/docs/models/index.mdxpackages/web/src/content/docs/docs/reference/ai-providers.mdpackages/web/src/content/docs/docs/reference/analysis-pipeline.mdpackages/web/src/content/docs/docs/reference/authentication.mdpackages/web/src/content/docs/docs/reference/configuration.md
| | `model-path` | Path to the model within the model repository | No | `.` | | ||
| | `model-ref` | Git ref (branch/tag) of the model repository | No | `main` | | ||
| | `model-format` | Architecture model format | No | `likec4` | | ||
| | `ai-provider` | AI provider (`gemini`, `openai`, or `anthropic`) | No | `anthropic` | |
There was a problem hiding this comment.
Inconsistency: Default AI provider mismatch with README.
The README states "Supported AI providers: Gemini (default)" (line 7), but this table shows anthropic as the default value for ai-provider. Since all workflow examples in this PR are being updated to explicitly set ai-provider: gemini, the table default should likely be updated to match.
📝 Proposed fix
-| `ai-provider` | AI provider (`gemini`, `openai`, or `anthropic`) | No | `anthropic` |
+| `ai-provider` | AI provider (`gemini`, `openai`, or `anthropic`) | No | `gemini` |🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@packages/web/src/content/docs/docs/integrations/github-actions.md` at line
76, The table entry for the `ai-provider` option is inconsistent with the README
and workflow examples; change the default value shown from `anthropic` to
`gemini` so `ai-provider`'s default matches the README and the updated
workflows, and scan other docs mentioning `ai-provider` to ensure all defaults
are consistently `gemini`.
Greptile SummaryThis PR updates the documentation to reflect the current recommended Gemini configuration and several new features across the action inputs, LikeC4 embeds, and Bitbucket authentication.
Confidence Score: 4/5Safe to merge with the understanding that the action.yml default issue is a follow-up fix rather than a blocker; all examples correctly show gemini. The action.yml default for ai-provider remains 'anthropic' while every example in this PR and the ai-providers.md page treat Gemini as the recommended default. A workflow author who skips the examples and relies only on the inputs table would silently get the experimental Anthropic provider. The rest of the changes are clean and consistent. github-actions.md and action.yml (not changed in this PR) — the action input default and application default disagree, and the Bitbucket ERODE_MODEL_REPO_TOKEN addition in authentication.md has no corresponding entry in the configuration.md platform tokens table. Important Files Changed
Flowchart%%{init: {'theme': 'neutral'}}%%
flowchart TD
A[User sets up GitHub Actions workflow] --> B{ai-provider specified?}
B -- Yes, gemini --> C[Uses Gemini - recommended]
B -- No / omitted --> D[action.yml default: anthropic]
D --> E[Experimental Anthropic provider used]
C --> F[Analysis runs with Gemini]
subgraph Docs
G[ai-providers.md: Gemini default]
H[configuration.md: ERODE_AI_PROVIDER default = gemini]
I[github-actions.md table: ai-provider default = anthropic]
J[action.yml: default = anthropic]
end
I -.matches.-> J
G -.conflicts with.-> J
H -.conflicts with.-> J
Reviews (1): Last reviewed commit: "docs: fix embedded architecture views" | Re-trigger Greptile |
| | `model-path` | Path to the model within the model repository | No | `.` | | ||
| | `model-ref` | Git ref (branch/tag) of the model repository | No | `main` | | ||
| | `model-format` | Architecture model format | No | `likec4` | | ||
| | `ai-provider` | AI provider (`gemini`, `openai`, or `anthropic`) | No | `anthropic` | |
There was a problem hiding this comment.
Action default still
anthropic while all examples show gemini
The ai-provider action input in action.yml still has default: 'anthropic', meaning any workflow that omits this field (following the table's "No / anthropic" row) silently uses the experimental Anthropic provider rather than Gemini. This contradicts both the application-level default documented in configuration.md (ERODE_AI_PROVIDER → gemini) and the ai-providers.md page which calls Gemini the default provider. Adding explicit ai-provider: gemini to every example is a good workaround, but the root fix is updating action.yml's default: from 'anthropic' to 'gemini' so the action input and the env-var default agree.
| ## Bitbucket (experimental) | ||
|
|
||
| `ERODE_BITBUCKET_TOKEN` handles all operations. There is no separate model-repo token. If the token contains `:` (e.g. `username:app_password`), Erode uses HTTP Basic auth; otherwise it uses Bearer auth. | ||
| `ERODE_BITBUCKET_TOKEN` handles source pull request operations. For external model repositories, the Bitbucket Pipelines entrypoint accepts `ERODE_MODEL_REPO_TOKEN` and falls back to `ERODE_BITBUCKET_TOKEN` when it is not set. If the token contains `:` (e.g. `username:app_password`), Erode uses HTTP Basic auth; otherwise it uses Bearer auth. |
There was a problem hiding this comment.
ERODE_MODEL_REPO_TOKEN for Bitbucket not in the configuration reference
This line introduces ERODE_MODEL_REPO_TOKEN as a new Bitbucket feature, but the Platform tokens table in configuration.md doesn't list it (it only lists ERODE_MODEL_REPO_PR_TOKEN for GitHub). Users consulting the configuration reference to set up Bitbucket model-repo separation would find no mention of this variable. Adding a row for ERODE_MODEL_REPO_TOKEN alongside the GitLab/Bitbucket entries would complete the reference.
Summary by CodeRabbit
ai-provider: geminiinput.ERODE_LIKEC4_FORMAT_AFTER_PATCHenvironment variable.