## Summary
Align the Go ingestion pipeline with Python's `image` → `img_id`
persistence semantics, and unify the chunk-id computation across all
paths.
### Changes
**1. Image upload at chunker stage **
- Add `ImageUploader` type and `DefaultImageUploader` in
`internal/ingestion/component/image_uploader.go` — the write-side
counterpart to `FetchBinary`, storing raw image bytes at `(bucket=kbID,
key=chunkID)`, no re-encoding.
- Add `uploadOneImage` — pure upload primitive (bytes in, `img_id` out),
does not touch chunk maps.
- Add `uploadChunkImages` / `uploadChunkImage` — caller-side helper:
decodes `image` from a chunk, uploads bytes, writes `ck["img_id"]`,
`delete(ck,"image")` , bounded by a process-wide semaphore (default 10,
env `MAX_CONCURRENT_MINIO`).
- Wire via `imageUploadDecorator` in `register.go`: every chunker runs
the upload pass at invocation time, writing `ck["id"]` before upload and
dropping image bytes right after — peak memory = single chunk image
lifetime.
**2. Unify chunk-id computation**
- Consolidate three separate id-computation paths (`component.ChunkID`,
`task.ChunkID`, inline `FormatUint` in API) into one:
`common.ChunkID(docID, text string)`, using `%016x` +
`xxhash.Sum64String(text+docID)` (matching Python `hexdigest()`).
- The chunker decorator writes `ck["id"]` via `common.ChunkID`; the
persist stage (`ProcessChunksForPipeline`) falls back to the same
function (`if !exists id`).
- The API AddChunk path now also calls `common.ChunkID` instead of the
divergent `FormatUint(xxhash.Sum64(...))` — fixing a pre-existing
inconsistency.
- Delete `internal/ingestion/component/chunk_id.go` and
`internal/ingestion/task/chunk_builder.go` (both were pure forwarding
shells).
**3. Preserve `img_id` (never deleted)**
- `img_id` is a persistent index field (Infinity, OB) and the only
consumer-side reference for image retrieval; it is NEVER removed from
the chunk map. Only `image` (raw data URL) is dropped after upload.
**4. PPT parser support**
Previously PPT parsing failed. Add support to parse.
### Key design decisions
| Decision | Choice |
|----------|--------|
| Upload timing | Chunker stage (not persist), so image bytes are
dropped immediately — bounds peak memory to one chunk image |
| Upload concurrency | Process-wide semaphore, default 10 (matches
Python `minio_limiter`), env `MAX_CONCURRENT_MINIO` |
| Image encoding | Store as-is, no JPEG re-encoding (unlike Python) |
| `img_id` format | `"<kb_id>-<chunk_id>"` — matches Python
task_executor path |
| id function | Single `common.ChunkID(docID, text)`, concatenation
`text+docID` inside hash (matching Python) |
| `removeInternalChunkFields` | Retains `delete(ck,"image")` as
defensive fallback for non-chunker paths |
### Files touched
| File | Change |
|------|--------|
| `internal/common/format.go` | Add `ChunkID(docID, text)` |
| `internal/common/format_test.go` | Add ChunkID golden-value test |
| `internal/ingestion/component/image_uploader.go` | Add `ImageUploader`
type + `DefaultImageUploader` |
| `internal/ingestion/component/chunker/image_upload.go` | Add
`uploadOneImage`, `uploadChunkImages`, `uploadChunkImage`,
`decodeChunkImage`, semaphore |
| `internal/ingestion/component/chunker/image_upload_test.go` | Tests:
upload/drop, skip, no-image, concurrency, missing-id error |
| `internal/ingestion/component/chunker/register.go` | Add
`imageUploadDecorator` (writes `ck["id"]`, runs upload) |
| `internal/ingestion/task/chunk_process.go` | Use `common.ChunkID` for
persist fallback |
| `internal/service/chunk/chunk.go` | Use `common.ChunkID` instead of
`FormatUint` |
| `internal/ingestion/component/chunk_id.go` | **Deleted** (moved to
`common`) |
| `internal/ingestion/task/chunk_builder.go` | **Deleted** (shell, no
callers left) |
| `internal/ingestion/task/chunk_builder_test.go` | **Deleted** (test
migrated to `common/format_test.go`) |
### Verification
```
bash build.sh --test ./internal/service/chunk/... ./internal/common/... ./internal/ingestion/component/... ./internal/ingestion/task/...
→ ok service/chunk / common / component / chunker / schema / task
```
### Summary
`agent/tools/github.py` indexes `response["items"]` directly after
`requests.get(...).json()`. the github search api returns `{"message":
...}` **without** an `items` key on realistic conditions — a rate limit
(403/429; this tool sends no auth token, so the unauthenticated ~10
req/min search limit is easy to hit) or an invalid query (422). that
raised `KeyError('items')`, which the tool's retry loop then surfaced to
the model as the opaque `"GitHub error: 'items'"` instead of the real
reason.
this guards the missing key and raises the api's actual `message` (with
a clear fallback) into the existing retry/`_ERROR` path, so the model
sees e.g. `"GitHub error: API rate limit exceeded ..."`. valid responses
are unchanged.
adds `test/unit_test/agent/tools/test_github_unit.py` covering the
rate-limit response (asserts the real message is surfaced, no
`KeyError`) and the normal result path.
Co-authored-by: Yaroslav98214 <diakovichyaroslav30@gmail.com>
Co-authored-by: Haruko386 <tryeverypossible@163.com>
## Summary
`Pipeline.Run` already accepts an `override_params` map (keyed by
`cpnID`, override-wins) that is merged into each component's params at
compile time via `canvas.WithOverrideParams`. `PatchDSL` performed a
byte-for-byte equivalent merge by baking `ParserConfig` into the DSL
text before compilation, so `runPipelineWithDSL` applied the same
`Doc.ParserConfig` twice with no effect.
## Change
- Delete the now-dead `PatchDSL` function
(`internal/ingestion/pipeline/pipeline.go`); it had a single call site.
- In `runPipelineWithDSL`, compile the DSL unchanged and pass the
extracted component params to `Run` as `override_params`.
## Behavioral note (important)
Previously `PatchDSL` was also the channel that delivered the tenant LLM
id into Extractor components: it baked `parserConfig` — after
`common.InjectExtractorLLMID(parserConfig, Tenant.LLMID)` had injected
`llm_id` — into the DSL. The old `Run` override_params re-cast
`Doc.ParserConfig` (which intentionally omits `llm_id`). So removing
`PatchDSL` while keeping that re-cast would have silently dropped LLM id
injection. This change passes the local `parserConfig` (the
LLMID-injected copy) to `Run`, preserving the prior behavior.
The DSL string returned by `runPipelineWithDSL` is only consumed by
`recordPipelineLog` (structure storage) and `ExtractPayload` (which
reads the terminal structure, not params), so returning the original DSL
is equivalent.
## Test plan
- `bash build.sh --test ./internal/ingestion/task/...
./internal/ingestion/pipeline/...` — both packages pass.
- `gofmt` clean; pre-commit hooks pass.
Co-authored-by: Claude <noreply@anthropic.com>
## What
The Japanese locale (`web/src/locales/ja.ts`) was missing translations
for **1703 of 2686 keys (~63%)**. Because i18next's `fallbackLng`
silently falls back to English for any missing key, most of the UI
stayed in English even after selecting 日本語 — e.g. the Settings > Model
Providers page, the Agent/Flow builder, dataset configuration panels,
and more.
This PR fills in all 1703 missing keys with Japanese translations.
Follow-up to #16468.
PR #16468 fixed the Python
(`api/db/joint_services/tenant_model_service.py
split_model_name`) and Go (`internal/service/model_service.go
parseModelName`)
parsers to right-anchor the '@'-split on the composite
"model_name@instance@provider" key, but the matching helper in the
front-end
(`web/src/utils/llm-util.ts parseModelValue`) was missed.
parseModelValue used `split('@')` (anchored on the first '@'), which
silently
mangles composite keys whose model_name itself contains '@' (e.g. LM
Studio
embedding IDs like `text-embedding-nomic-embed-text-v1.5@q8_0`):
text-embedding-nomic-embed-text-v1.5@q8_0@lmstudio@LM-Studio
became
model_name: text-embedding-nomic-embed-text-v1.5
model_instance: q8_0@lmstudio
model_provider: LM-Studio
`PATCH /api/v1/models/default` then sent those to the server, which
returned
HTTP 200 with body `{"code": 102, "message": "Instance 'q8_0@lmstudio'
not
found for provider 'LM-Studio'"}`. The UI swallowed the body-level error
code, so the failure was silent: no toast, the Embedding field stayed
empty.
This change makes parseModelValue do a right-anchored split that mirrors
the
Python `split_model_name` `rsplit('@', 2)` exactly:
- 3-part form: model_name@instance@provider -> same three fields.
- 2-part form: model_name@provider -> model_instance defaults to
"default"
(matching Python).
- 4+-part form (embedded '@' in model name): last two fields are
anchored
as instance and provider, everything to the left is the bare model name.
`parseModelUuid` is updated to the same right-anchored split so the
factoryId portion of `model_name@factory_id[#instance]` keeps the last
'@'
as the separator even when the model name itself contains '@'.
Adds jest cases under `web/src/utils/tests/llm-util.test.ts` (12 total)
covering the plain 3-part, 2-part, 4-part, multi-'@' name,
buildModelValue
round-trip and parseModelUuid variants.
Note for reviewers: the front-end is bundled into the published docker
image, so a release containing this fix will need `docker compose build
--no-cache ragflow-cpu` (or equivalent) for it to be visible to end
users.
Refs #16468, #16467
