Docs/remove byok billing section#844
Conversation
Co-authored-by: factory-droid[bot] <138933559+factory-droid[bot]@users.noreply.github.com>
Co-authored-by: factory-droid[bot] <138933559+factory-droid[bot]@users.noreply.github.com>
…analytics API Co-authored-by: factory-droid[bot] <138933559+factory-droid[bot]@users.noreply.github.com>
|
Droid finished @francesca-factory's task —— View job |
![factory-droid[bot]](https://avatars.githubusercontent.com/in/358017?s=60&v=4){kind=small}
| return body | ||
|
|
||
| # Split content into code blocks and non-code segments | ||
| parts = re.split(r'(```[\s\S]*?```)', body) |
There was a problem hiding this comment.
[P1] Translation no longer deterministically preserves inline code/MDX/link structure
translate_body only protects fenced code blocks, so inline code spans (backticks), JSX/MDX expressions ({...}), and link/image syntax are now sent to the model without placeholders/AST protection. Even small model drift can corrupt MDX and break the docs build; consider reintroducing deterministic preservation for inline code + JSX/MDX constructs + link destinations (or parsing MDX and translating only text nodes/attribute values).
| - MDX component tags and structure | ||
| - Line breaks and paragraph structure | ||
|
|
||
| 4. For links pointing to English docs paths, update them to Japanese paths: |
There was a problem hiding this comment.
[P2] Prompt has conflicting instructions about URLs/paths vs internal link rewriting
The prose prompt says “Do NOT translate URLs and file paths” but also asks the model to rewrite internal docs paths (e.g. /cli/... → /jp/cli/...). This conflict can produce inconsistent link handling across chunks; it’s more reliable to keep the prompt strict (“never change URLs”) and implement any required internal-link rewriting deterministically in code.
| - **Without `group_by`**: Returns one row per day with nested breakdowns (e.g., `by_model`, `by_tool`, `daily_active_users_by_client`). Use this when you want all dimensions in a single response. | ||
| - **Without `group_by`**: Returns one row per day with nested breakdowns (e.g., `by_tool`, `daily_active_users_by_client`). Use this when you want all dimensions in a single response. | ||
|
|
||
| - **With `group_by`**: Flattens one of those nested arrays into separate rows. Each row has a `group_key` field identifying the dimension value. Use this when piping data into tools that expect flat rows (spreadsheets, BI tools, time-series databases). |
There was a problem hiding this comment.
[P2] group_by behavior description contradicts the /activity example
The doc says group_by “Flattens one of those nested arrays,” but immediately below notes /activity returns daily_active_users_by_client as an object that can also be grouped into rows. Reword to “nested breakdown field (array or object map)” (or similar) to avoid misleading consumers about the response shape.
There was a problem hiding this comment.
Docs changes look straightforward, but the translation script change increases the risk of corrupting MDX since only fenced code blocks are preserved and everything else relies on model compliance. There’s also a small inconsistency in the Analytics API group_by description vs the /activity example that’s worth correcting.
2 participants
No description provided.