Repository navigation
Fixes and improvements - #428
Open
Chenglong Wang (Chenglong-MS) wants to merge 76 commits into
Open
Chenglong Wang (Chenglong-MS) wants to merge 76 commits into
Chenglong Wang (Chenglong-MS) wants to merge 76 commits into
Conversation
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
Add orcarouter to the built-in providers so Data Formulator users can connect the OrcaRouter AI gateway from the model picker and via ORCAROUTER_* environment variables, mirroring the existing ollama/openai wiring. The client routes through LiteLLM's openai provider against the OrcaRouter base URL, preserving the orcarouter/ model namespace.
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
…ovider feat: add OrcaRouter as a built-in model provider
Adds full Hindi translations for all 10 i18n domain files, following the translation guide's rules (chart type identifiers, encoding channel keys, and other computation-bound values stay in English). Registers hi in the locale index/i18next resources and language switcher labels, and documents AVAILABLE_LANGUAGES=en,zh,hi as opt-in in .env.template. Relates to #337.
folderPathPlaceholder illustrates filesystem path syntax; translating the segment words broke the example and was inconsistent with how every other path/URL placeholder in the app is left untranslated.
Add Hindi (hi) locale for frontend i18n
Comment on lines
+823
to
+826
| '<!doctype html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width">' | ||
| '<title>OpenRouter</title></head><body><p>' + message + '</p>' | ||
| '<a href="/">Return to Data Formulator</a>' | ||
| f'<script nonce="{nonce}">{script}</script></body></html>', |
| creator_id = key_info.get("creator_user_id") | ||
| connection = { | ||
| "creator_user_id": creator_id if isinstance(creator_id, str) else None, | ||
| "settings_url": "https://openrouter.ai/keys/" + hashlib.sha256(config["api_key"].encode()).hexdigest(), |
The wrapper embeds user-supplied code, so on a system whose preferred encoding is not UTF-8 (cp1252 on Windows) writing it raised UnicodeEncodeError for any non-ASCII character, and silently transcoded the rest -- the container reads run.py as UTF-8 either way. Matches the explicit encoding already used elsewhere in the package, e.g. reasoning_log.py and analyst/skills/__init__.py.
The constant table listed DEFAULT_ROW_LIMIT_EPHEMERAL in dfSlice.tsx. That identifier does not exist anywhere in the repository, and there is no separate ephemeral row limit in the frontend.
…hemeral-row docs(dev-guides): drop the nonexistent DEFAULT_ROW_LIMIT_EPHEMERAL row
fix(sandbox): write the docker wrapper script as utf-8
Translate all 10 frontend locale domains into Bahasa Indonesia (1860 keys, full parity with en). Register the bundle in locales/index.ts and i18n/index.ts, add the 'Bahasa Indonesia' language switcher label, and cover key parity with a unit test.
Add 36 missing Hindi keys and 5 missing Simplified Chinese keys introduced by newer features (semantic model sidebar, connector groups, chart quick actions, virtual sources). All four UI locales now share the same 1860-key set.
Add Indonesian (id) locale + backfill missing hi/zh keys to full parity
Translate all 10 frontend locale domains into Japanese in polite desu/masu form (1860 keys, full parity with en). Register the bundle in locales/index.ts and i18n/index.ts, and cover key parity with a unit test. The language switcher label and backend language support for Japanese already existed, so no changes were needed there.
| paths = filesystem['allowWrite'] | ||
| if (not isinstance(paths, list) or len(paths) > 64 | ||
| or any(not isinstance(path, str) or not path or len(path) > 2000 or '\0' in path | ||
| or not Path(path).expanduser().is_absolute() for path in paths)): |
| home = Path.home().resolve() | ||
| protected = configuration_path().parent.resolve() | ||
| for path in paths: | ||
| resolved = Path(path).expanduser().resolve() |
| actor, saved['revision'], ','.join(changed_sections)) | ||
| return json_ok(snapshot()) | ||
| except ConfigurationConflict as exc: | ||
| return {'status': 'error', 'error': {'code': 'INVALID_REQUEST', 'message': str(exc), 'retry': False}}, 409 |
Feat/locale ja
Calling ensure_table_keys() at the base call sites does not cover loaders that override the tree builders. SupersetLoader overrides both list_tables_tree() and search_catalog() and constructs its nodes by hand, and the clickhouse, mysql and postgresql search_catalog implementations bypass the base method too. Move the backfill into _tables_to_catalog_tree(), which every tree builder apart from the Superset overrides funnels through, and apply the key on the Superset side separately. That path keys on the dataset uuid in list_tables(), so the uuid is now carried through ls(), _build_dashboard_group_metadata() and _dataset_meta() to keep both paths producing the same key.
| * Open a saved session, replacing the current one. Resolves true when it opened; | ||
| * failures are reported to the user and resolve false. | ||
| */ | ||
| export const openSession = (sessionId: string, displayName?: string, options: { saveCurrent?: boolean } = {}) => |
fix(data-loader): backfill table_key in tree and search responses
Add an html_app analyst skill whose write_html_app action saves an agent-managed .html workspace file with a manifest of the tables it may read. Revisions go through edit_file. The canvas renders .html files in a double iframe: a same-origin container with frame-src 'none' (blocks navigation exfiltration) holding an opaque-origin sandbox="allow-scripts" app frame with a CSP that blocks network access. An inlined DF runtime provides DF.query/DF.table (validated bridge to sample-table, manifest tables only), DF.chart (Vega-Lite), and theme variables. Remove the unused Chartifact dialog and its i18n keys. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
9 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This pull request introduces several improvements and refinements to the Data Formulator agent and context handling, focusing on clearer terminology, enhanced connector and skill management, and dependency updates. The most important changes are summarized below.
Agent and Context Terminology Improvements:
Connector Availability and Summary Enhancements:
Skill Preloading and System Prompt Improvements:
Dependency Updates:
vite,dompurify,echarts,js-yaml, and new resolutions forpostcss,esbuild, andtmp. [1] [2] [3]Documentation Cleanup:
docs/desktop-portable.mdfile has been removed.loops/model-evaluation/plan.mdhas been deleted.