Unifies the Go TokenChunker merge path on a single `mergeUnits` core and
fixes coordinate-tag drift in the Python JSON merge at `overlap > 0`.
Rebased on top of #17979 (delimiter_mode convergence).
Converge `TokenChunker.delimiter_mode` from three values (`token_size`,
`delimiter`, `one`) to two (`delimiter`, `one`). The unified `delimiter`
mode now carries the old `token_size` semantics: when no active
(backtick) delimiter is present, text/JSON chunks are merged up to
`chunk_token_size`; when a backtick delimiter is present, the text is
split by it and not merged. `one` continues to be handled by the
separate `OneChunker`.
Restore the deduplication that was dropped when #17926 was merged.
`CompileDelimiterPatternList` now keeps a `seen` set and collapses
equivalent active entries (both backtick-inner and bare) into a single
alternation. This PR also removes the dead code that the re-review
surfaced.
`compileChildrenPattern` re-implemented the delimiter-list compile
inline with two divergences from the shared
`CompileDelimiterListPattern`:
- It never stripped backticks, so a backtick-wrapped
`children_delimiter` like `` `###` `` matched the **literal wrapped
token** rather than the inner `###`.
- It sorted by **byte length** instead of rune count (`sortSlice`), so
multi-byte delimiters could be ordered incorrectly and a longer
delimiter could fail to win over a shorter prefix.
## Summary
- Remove `splitOversizedUnit`, `splitAtomByTokenBudget` and `atomRE`
from `internal/ingestion/component/chunker/token.go`.
- Delete `split_oversized_guard_test.go` (added by #17740), which
guarded the removed atom-split behaviour.
- Drop the now-unused `wordCount`/`charCount` helpers from
`token_strict_cap_test.go`.
- Add `TestMergeByTokenSize_OversizedUnitStaysWhole` to pin the #17799
contract invariant (over-budget unit stays whole, never atom-split) on
the **text path**. The JSON path is already covered by
`TestMergeByTokenSizeFromJSON_OversizedUnitStaysWhole`.
## Why
The production merge path (`mergeByTokenSize` /
`mergeByTokenSizeFromJSON`) keeps over-budget units whole and relies on
the embedding/rerank layer to truncate them, per the TokenChunker
contract (#17799: remove atom-split, no hard_cap). The deleted helpers
implemented the opposite behaviour and had **no production caller**, so
they contradicted the contract and misled readers into thinking
atom-split was active.
## Parser vs chunker layering
Python's `_split_oversized_unit` lives at the **parser layer**
(pre-split before `naive_merge`), not in the chunker. Go's parser
backends are currently skeletons, so there is no parser-side equivalent
yet; if added later it belongs in `internal/parser/parser/*`, not the
chunker.
## Test plan
`bash build.sh --test ./internal/ingestion/component/chunker/...`
passes; the new text-path test passes and the orphaned atom-split tests
are gone.
## Changes
- 3 files changed, 32 insertions(+), 250 deletions(-)
Ports the dataset knowledge compilation (wiki/graph/tree/mindmap) to the
Go scheduler with a status contract, aligns wiki storage/retrieval with
Python, and sizes prompts by content_length.
The Go `TokenChunker` text/markdown path (`mergeByTokenSize`)
unconditionally
called `splitOversizedUnit` on any unit that exceeded
`chunk_token_size`,
emitting Go-only sub-chunks. Python's `naive_merge`
(`_merge_paragraph_groups`,
`rag/nlp/__init__.py`) never atom-splits an oversize unit under either
`OVER_CAP` or `UNDER_CAP`: a paragraph larger than the budget becomes
its own
standalone chunk and the model layer truncates it later.
This aligns the text/markdown path with the **structured JSON path**
(`invokeJSONPayload` → `mergeByTokenSizeFromJSON(...,
subSplitOversize=false)`,
#17739). It completes the OVER_CAP alignment started in #17835.
Go's `TokenChunker` kept the captured delimiter glued to the preceding
segment on **both** the primary (`chunkFromItem`) and secondary
(`children_delimiters`) split paths, while Python's reference
`token_chunker` drops it via `_split_text_by_pattern`
(`token_chunker.py:79-93`, used by both `_build_json_chunks` and
`_split_chunk_docs_by_children`). The divergence leaked the delimiter
into every emitted chunk's `text`.
## Summary
Follow-up to #17835 (merged). The OVER_CAP / UNDER_CAP merge strategy
was threaded through `mergeDecision` and `mergeByTokenSizeFromJSON` as
an inlined `allowBoundaryOverflow bool` derived from `!c.param.UnderCap`
at three call sites. This replaces that with a named
`schema.MergeStrategy` enum.
## Why
- The `!c.param.UnderCap` inversion was hand-written in three places, so
a future strategy addition could silently drift between the JSON path
(`invokeTextPayload` / `invokeJSONPayload`) and the text path
(`mergeByTokenSize`) — no compile error, and the existing tests don't
cover all three sites with both strategies.
- The strategy concept was never named; `allowBoundaryOverflow` (true =
OVER_CAP) is a double-negation of `UnderCap` and reads opaquely at the
5th positional argument.
## What changed
- Add `schema.MergeStrategy` (`MergeOverCap` / `MergeUnderCap`)
mirroring Python's `rag/nlp/__init__.py` `MergeStrategy`, so Go and
Python stay on the same vocabulary.
- Expose `TokenChunkerParam.MergeStrategy()` derived from the
wire-facing `UnderCap bool` (existing `"under_cap"` configs keep working
— no schema break).
- `mergeDecision` and `mergeByTokenSizeFromJSON` now take
`schema.MergeStrategy` instead of `allowBoundaryOverflow bool`; the
three call sites pass `c.param.MergeStrategy()` (no `!`).
- Tests updated to pass the enum; added a guard test for the `UnderCap`
-> `MergeStrategy` mapping and an end-to-end test for UNDER_CAP on the
JSON path.
No behavior change: default remains OVER_CAP, `under_cap=true` still
selects UNDER_CAP.
## Test plan
`bash build.sh --test ./internal/ingestion/component/chunker/...
./internal/ingestion/component/schema/...` — all green, including
`TestMergeByTokenSizeFromJSON_UnderCapNoOverflow`,
`TestMergeByTokenSize_UnderCapNoOverflow`,
`TestInvokeJSONPayload_UnderCapEndToEnd`, and
`TestTokenChunkerParamMergeStrategy`.
## Related issues
- Relates to #17835 — wired UNDER_CAP as a tested merge-strategy seam
(merged)
- Relates to #17799 — contract doc for token-chunker cap/delimiter
alignment
- Relates to #17808 — related chunker alignment work
---------
Co-authored-by: CodeBuddy <noreply@cnb.cool>
## Summary
Align Go `splitOversizedUnitWith` with Python
`rag/nlp._split_oversized_unit` so the whitespace-atom sub-split
produces byte-identical chunk boundaries.
### Root cause of the divergence
cl100k token counting is **not additive across whitespace joins**
(`token(a)+token(b) != token(a+b)`). Go previously used the exact
joined-string fit check `countFn(current+atom) > budget`, while Python
accumulates a running sum `current_tokens + a_tokens > budget`. The two
formulas disagree by one atom at the boundary, so Go and Python emitted
the same chunk *count* but shifted *text*.
### Changes
- `splitOversizedUnitWith` (`token.go`): replace the exact joined-string
fit check with the running-sum check (mirroring Python's
`current_tokens` accumulator), and after a flush keep the overflow
whitespace atom (`current += atom`) instead of dropping it.
- `token_strict_cap_test.go`: relax
`TestMergeByTokenSizeFromJSON_OversizedUnitIsSubSplit` to allow the same
cl100k non-additive +1 overshoot Python exhibits (the invariant — an
oversized unit is sub-split, not collapsed — is preserved).
### Test plan
`bash build.sh --test ./internal/ingestion/component/chunker/...` —
green.
## Note
Test infrastructure for this change (golden parity harness,
`split_oversized_test.go`, `testdata/parity/**`, `known_diffs.json`,
`capture_golden.py`/`live_chunk.py`, and the `go-cmp` dependency
promotion) is split into a separate, stacked PR #17735 so this PR stays
minimal (production code only).
This PR is **independent of #17712** (the offline BPE loader). It is
based on `upstream/main` and contains only this change; no BPE-loader
code is included.
Co-authored-by: CodeBuddy <noreply@cnb.cool>
## Summary
Six sites used to read the same `parser_config.delimiter` field with
divergent grammars:
- `rag.nlp.get_delimiters` (PDF/DOCX/HTML/EPUB/JSON/CSV/XLSX/email/book)
- `rag.nlp.naive_merge` (custom-delimiter branch)
- `rag.nlp.naive_merge_with_images`
- `rag.nlp._build_cks`
- `deepdoc.parser.txt_parser.parser_txt` (.txt, code)
-
`deepdoc.parser.markdown_parser.MarkdownElementExtractor.get_delimiters`
The six implementations disagreed on bare-vs-wrapped chars, dedupe, sort
order, CRLF normalization, and `re.I` (#17384). The shipped default ``
`\n!?;。；！？` `` was a no-op for `.md` because the markdown path only
matched backtick-wrapped tokens.
## Changes
- **new:** `rag/nlp/delim.py` with `parse_delimiter_field` and
`compile_delimiter_pattern`. Single source of truth. CRLF normalization
at the top; longest-first stable sort; insertion-ordered dedupe; no
`re.I`.
- **refactor:** all six call sites delegate to the helper.
- `rag/nlp/__init__.py::get_delimiters` becomes a thin shim.
- `deepdoc/parser/txt_parser.py::parser_txt` drops the
`[encode/decode/unicode_escape]` round-trip.
- `deepdoc/parser/markdown_parser.py::get_delimiters` honors bare chars
(fixes [1]).
- **tests:** `test/unit_test/rag/test_delim.py` (85 tests) — helper,
acceptance table, frontend parity, static guard against re-inlining.
- **tests:** `test/unit_test/rag/test_delimiter_case_sensitive.py` (from
#17386) updated to retarget the static check at the new helper +
AST-based broader scan.
## Acceptance criteria
- All six sites produce the same regex pattern for the same input.
- Shipped default keeps working for `.txt` / `.pdf` / `.docx`.
- Shipped default for `.md` now splits (was a silent no-op).
- Tooltip example `` `\n##;` `` produces three effective delimiters
regardless of file type.
- Bare whitespace inputs split on every occurrence.
- Backtick-wrapped whitespace splits only on the exact N-char sequence.
- CRLF-line-ending documents split identically to LF-line-ending
documents.
- 123 tests pass (85 new + 38 existing).
## Rebase protocol
As #17385 and #17386 evolve, this branch will be rebased on top. The
only overlap between this PR's diff and the other two is
`test_delimiter_case_sensitive.py`, where #17383 modifies the static
check to point at the new helper location.
---------
Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
Fixes#17202 (and complements #12109).
## Problem
`RAGFlowTxtParser.parser_txt` (`deepdoc/parser/txt_parser.py:36-47`) and
`rag.nlp.naive_merge` (`rag/nlp/__init__.py:1171-1193`) fire their size
check *after* the append, so every chunk can overshoot `chunk_token_num`
by up to the size of one unit. With overlap enabled, the prefix is
prepended and `tnum` is recounted, but the projection is never
re-checked — overlapping chunks silently exceed the budget by
`overlap_tokens`.
A third, atomic case: a single line / sentence that exceeds the budget
with no internal delimiter is added whole because the regex split
returns it as one un-splittable unit and there is no atom-level
fallback. `RAGFlowHtmlParser.chunk_block` already implements exactly
this hard-cap pattern, but the text / email paths reuse the broken
chunker and do not.
Measured on a live dataset (336 `.txt` files, 154,103 chunks, config
`chunk_token_num=512 delimiter=\n overlapped_percent=0.1`): 56.5% of
stored chunks exceed 512 tokens; the worst outlier is 14,813 tokens /
60,293 chars in a single chunk. Symptom downstream: rerank failures on
the >2048-token outliers (ref. #12109) and silent embedding truncation
on every oversize chunk.
## Fix
Mirror the proven pattern in `RAGFlowHtmlParser.chunk_block`:
1. **Proactive projected-total check** in `TxtParser.parser_txt` and in
`naive_merge.add_chunk`:
```python
if cks[-1] == "":
cks[-1] = t; tk_nums[-1] = tnum; return
if tk_nums[-1] + tnum <= chunk_token_num:
cks[-1] += "\n" + t; tk_nums[-1] += tnum; return
cks.append(t); tk_nums.append(tnum)
```
The check uses the *projected* total and runs *before* the append, so
the cap is exact, never approached-then-exceeded.
2. **Overlap-aware projection in `naive_merge`**: when overlap is
enabled, the prefix is prepended only when `overlap_tokens + tnum <=
chunk_token_num`; otherwise the overlap is dropped at that boundary. The
naive_merge-with-images mirror gets the same treatment. Custom-delimiter
behaviour is preserved per the existing test suite.
3. **Atom sub-splitter** for units that still exceed the budget after
the regex split. Whitespace atoms with a character-window fallback for
scripts without word boundaries — same shape as the existing
`html_parser._split_oversized_block`, so behaviour matches for HTML vs
`.txt` vs PDF atomic-oversize.
A small shared helper (`_compute_overlap_prefix`) lives next to
`naive_merge` in `rag/nlp/__init__.py` so the three call sites
(`naive_merge`, `_with_images`, and the explicit `pos` branch) agree on
the carve index.
## Result on the dataset above
| | Before | After |
|---|---|---|
| Chunks > 512 tokens | 56.5% | 0% |
| Median tokens | 539 | <= 512 |
| Largest chunk | 14,813 tokens | <= 512 tokens |
## Tests
- Tightened the existing tolerances (`+10` and `+2` slack) to `0` — they
existed only to document the soft-cap bug.
- Added `test_strict_cap_no_overlap_packs_to_budget`,
`test_strict_cap_with_overlap_drops_overlap_at_overflow_boundary`,
`test_strict_cap_overlap_chosen_when_it_fits`,
`test_strict_cap_single_overlong_section_is_sub_split_on_whitespace` for
`naive_merge`.
- Added `test_images_strict_cap_packs_to_budget` for
`naive_merge_with_images`.
- New `test/unit_test/deepdoc/parser/test_txt_parser.py` covers
`parser_txt` strict cap and atom sub-split. Uses the same path-loading
pattern as the existing `test_html_parser.py` to avoid pulling the deep
import chain into a test-time-only venv.
All 22 unit tests pass on the host venv:
```
test_naive_merge.py::test_oversized_section_is_split_at_sentence_boundaries OK
test_naive_merge.py::test_small_sections_are_merged_not_oversplit OK
test_naive_merge.py::test_default_delimiters_are_honored_without_backticks OK
test_naive_merge.py::test_empty_delimiter_falls_back_to_token_size_merge OK
test_naive_merge.py::test_overlap_prefix_is_counted_in_token_budget OK
test_naive_merge.py::test_custom_delimiter_ignores_chunk_size OK
test_naive_merge.py::test_custom_delimiter_does_not_size_merge OK
test_naive_merge.py::test_images_oversized_section_is_split OK
test_naive_merge.py::test_images_custom_delimiter_preserved OK
test_naive_merge.py::test_images_plain_string_input OK
test_naive_merge.py::test_images_mismatched_lengths_returns_empty OK
test_naive_merge.py::test_images_shared_lazyimage_not_stacked_… OK
test_naive_merge.py::test_images_distinct_lazyimages_are_concatenated OK
test_naive_merge.py::test_strict_cap_no_overlap_packs_to_budget OK
test_naive_merge.py::test_strict_cap_with_overlap_drops_… OK
test_naive_merge.py::test_strict_cap_single_overlong_section_… OK
test_naive_merge.py::test_strict_cap_overlap_chosen_when_it_fits OK
test_naive_merge.py::test_images_strict_cap_packs_to_budget OK
test_txt_parser.py::test_no_overshoot_when_packing_short_lines OK
test_txt_parser.py::test_no_overshoot_at_chunk_boundary OK
test_txt_parser.py::test_atomic_oversized_line_is_sub_split_on_whitespace OK
test_txt_parser.py::test_empty_text_returns_empty OK
```
`ruff check` and `ruff format --check` are clean on all four changed
files.
## Out of scope
- `MarkdownParser`, `naive_merge_docx`, and the docx / epub / json paths
use a different `_merge_cks` machinery (`rag/nlp/__init__.py:1574`) that
already enforces the budget. They are unchanged.
- The `chunk_block` call sites in `deepdoc/parser/html_parser.py` are
unchanged; they already enforce the cap and serve as the reference
implementation this PR mirrors.
Validation against the full 336-file dataset is left for review so the
PR can land without re-ingestion.
---------
Co-authored-by: skbs-eng <skbs-eng@users.noreply.github.com>
Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
## Summary
Closes two Chunker migration gaps documented in
`docs/migration_python_go_diff.md`
(diffs **2.6** and **1.7**), improving parity with the Python ingestion
pipeline.
This is the code portion of commit `261e1fd0b` on branch
`fix/batch-3-4`; the
migration document itself is tracked separately (untracked in this
commit).
### Diff 2.6 — `overlapped_percent` missing Python normalization
Python's `normalize_overlapped_percent` (`common/float_utils.py:50-58`)
accepts a
`[0,1)` fraction (the flow-canvas UI validates `[0,1)`), multiplies it
by 100,
`int()`-truncates, and clamps to `[0,90]`. Go previously only accepted a
raw
`[0,90]` percentage and **rejected** out-of-range input, so a Python
config
passing `0.1` (meaning 10%) silently produced ~0% overlap.
Added `normalizeOverlappedPercent`
(`internal/ingestion/component/chunker/common.go`) mirroring the Python
helper:
- parses numbers and numeric strings (mirrors Python `float()`; bad /
`NaN` / `Inf` → `0`),
- `0 < v < 1 → v *= 100`,
- `int()` truncation,
- clamps to `[0,90]`.
Wired into `tokenChunkerParam.Update` (`token.go`); the merge math
(`token.go:705`, `(100-x)/100`) already matched Python.
`TokenChunkerParam.Validate`
(`schema/chunker.go`) now only guards direct struct construction.
Tests: `TestNormalizeOverlappedPercent`, extended
`TestTokenChunker_NewAcceptsPythonOverlappedRange` (adds fraction/clamp
inputs),
new `TestTokenChunker_NormalizesOverlappedPercent`. The existing
reject-test cases
for `<0` / `>90` were removed because they are now normalized/clamped
(Python parity).
### Diff 1.7 — missing `BULLET_PATTERN` title-level fallback
`resolveTitleLevels` (`title.go`) now applies a 4th-level fallback: when
outline +
regex + layout all yield body level, `bulletsCategory` selects the
best-matching
bullet-pattern group (Chinese legal / numbering / Chinese numbering /
English legal
— mirroring `rag/nlp/__init__.py:258-320`) and assigns structural
levels. Guarded by
`allBodyLevel` so it never overrides an existing outline/regex level.
Tests: `TestResolveTitleLevels_BulletFallback` (4 subtests).
### Incidental test adjustments included in the commit
- `token_batch1_test.go`: overlap input changed `0.3` → `30.0` to
reflect the
post-normalization 30% semantics.
- `real_consumer_test.go`: updated `LoadFromIngestionTask(task)` →
`LoadFromIngestionTask(ctx, task)` for the new context-first signature.
## Verification
`bash build.sh --test ./internal/ingestion/component/...` — chunker +
schema suites
pass, no regression (CGO build).
## Migration doc reference
`docs/migration_python_go_diff.md` §Chunker 1.7 and 2.6 are marked
**Fixed** for
these changes.
## Summary
Aligns the Go ingestion pipeline with the Python implementation, closing
several behavioral gaps found during the Python→Go migration (tracked in
`docs/migration_python_go_diff.md`). Covers parser/media dispatch
alignment, the PDF coordinate-chain (preview images, outline→title,
chunk coordinate finalization), and the Chunker Token/QA batches below.
Commits are grouped as follows.
### 1. Fix parser params (c524f450e)
Fixes parser/media wiring and several dispatch gaps:
- **docx/pdf vision dispatch**: correct parameter handling and VLM
invocation.
- **markdown vision (diff 2.5)**: also enhance items whose
`doc_type_kwd` is `table`, not only `image` (parser/utils.py:181).
- **media audio (diff 2.11)**: when `output_format` is `json`, carry the
ASR transcription as a JSON item instead of only the `Text` field (the
Invoke switch had no `json` branch and dropped it).
- **email (diff 2.2)**: default `output_format` is `json`
(parser.py:212), not `text`.
- **tokenizer**: handle empty/whitespace-only names; trim before
embedding.
- **extractor**: tag-matching parameter wiring.
- **split**: keyword-split regex now covers CJK/English separators.
- **parser.go**: parser-param plumbing.
### 2. fix parser gap (373537da1)
Image dispatch now mirrors `rag/app/picture.py:chunk()`:
- Always OCR the image (PaddleOCR or local ONNX).
- When OCR text is short, also call VLM (`describe`) and combine `OCR +
VLM` text.
- Emits a **structured JSON item** carrying the image data-URI and
`doc_type_kwd:"image"`, instead of a bare `Text` string. This fixes the
payload being rejected downstream by OneChunker/TokenChunker (JSON=nil).
### 3. PDF coordinate-chain fixes (55367a820, 727f8167c)
Closes three items from the migration tracker in the
chunker/tokenizer/task layer:
- **(Chunker-1.3) `restore_pdf_text_previews`** — `needsCrop` now also
returns true for `text` chunks that carry PDF positions
(`pdfcrop_cgo.go`), so text blocks get a rendered preview image uploaded
to storage via `imageUploadDecorator`/`ChunkImageUploader`, matching
Python `restore_pdf_text_previews` + `image2id`.
- **(Chunker-1.5) PDF outline → title levels** — `title.go` adds
`outlineSimilarity` (rune-bigram Jaccard, mirroring
`common.py:_outline_similarity`), `resolveOutlineLevels` (matches text
lines to outline entries at similarity > 0.8, with a sparse guard
`len(outline)/len(records) <= 0.03`), and `outlineFromInputs` (reads
`file.outline`). Wired into `newLevelContext` in both `group.go` and
`hierarchy.go`; falls back to the title-shape heuristic when no outline
is present.
- **(Tokenizer-(T)1) `finalize_pdf_chunk`** — the coordinate →
`position_int`/`page_num_int`/`top_int` conversion is owned by the task
layer (`processChunkPositions`→`AddPositions`), which runs *after* the
tokenizer and consumes the tokenizer-owned fields. The tokenizer only
preserves the raw `positions`/`_pdf_positions` (no duplicate
conversion), pinned by `TestChunkDocsToMaps_PreservesPDFPositions`.
### 4. Integration test made environment-free
(`internal/ingestion/task/pipeline_real_integration_test.go`)
- Removed the `//go:build integration` tag so the contract tests run
under the default `build.sh --test` (which does not pass `-tags
integration`).
- External dependencies replaced with in-memory substitutes so no
MySQL/MinIO/ES is required:
- MySQL → on-disk sqlite (`glebarez/sqlite`) with the needed tables
auto-migrated.
- MinIO → `storage.NewMemoryStorage()`.
- Elasticsearch → chunks captured via `WithInsertFunc` instead of
`engine.InsertChunks`/`Search`.
- `requireTokenizerPool` still skips gracefully when the native
tokenizer pool is unavailable; `WithLogCreateFunc(noop)` avoids
depending on the operation-log table.
- Added `taskChunkFieldEqualsStr` to tolerate `kb_id` being a
`[]string`/`[]any` in the raw chunk payload (the search engine flattens
it to a string on read).
### 5. TokenChunker alignment — Batch 1
(`internal/ingestion/component/chunker/token.go`)
Closes four Chunker items from the migration tracker:
- **(Chunker-2.1) sentence delimiter** — the boundary regex now also
breaks on ASCII `!`/`?`. Extracted to a package-level `var
sentenceDelimiter` and used in `mergeByTokenSize`, matching Python's
full delimiter set.
- **(Chunker-2.2) overlap tag leakage** — when a new chunk starts, its
overlap prefix is taken from the previous chunk *after* `removeTag`, in
both the text path (`mergeByTokenSize`) and the JSON path
(`mergeByTokenSizeFromJSON`). Parser tags (`@@…##`) no longer leak into
the overlap region (mirrors `nlp/__init__.py:1181`).
- **(Chunker-2.11) empty-text merge** — merging a non-empty chunk into
an empty previous chunk now assigns the text directly instead of being
skipped (`mergeByTokenSizeFromJSON`), mirroring
`token_chunker.py:236-239`.
- **(Chunker-2.4) overlap token counting** —
`takeFromEnd`/`takeFromStart` now count tokens exactly via `tokenizeStr`
instead of the 4-bytes/token heuristic, fixing over-counting for CJK
text.
### 6. QA Chunker alignment — Batch 2
(`internal/ingestion/component/chunker/qa.go` + `schema`)
Closes three Chunker items from the migration tracker:
- **(Chunker-2.13) default language** — an empty `lang` now defaults to
Chinese prefixes (`问题：`/`回答：`) instead of English, matching `qa.py:299`.
- **(Chunker-2.12) `rmQAPrefix` regex** — the separator is changed to
`[\t:： ]+` (one-or-more), matching `qa.py:241`, so multiple separators
(e.g. `Q:: answer`) are fully stripped.
- **(Chunker-1.8 QA) missing chunk fields** — QA chunks now preserve:
- `top_int` — the source row/record index, threaded through the
tab/csv/markdown extractors (mirrors `qa.py` `beAdoc(..., row_num=i)`);
- `image` + `doc_type_kwd:"image"`;
- `_pdf_positions` / `positions` carried from the upstream JSON item.
`schema.ChunkDoc` gains a `TopInt []int` field (serialized as `top_int`,
registered in `UnmarshalJSON`). Note: the Tag/Table/Presentation/One
chunker field gaps under 1.8 remain pending.
## Test plan
- Added/updated unit tests: `pdfcrop_cgo_test.go` (`TestNeedsCrop`,
`TestRestorePDFTextPreview`), `title_test.go`
(`TestResolveOutlineLevels`, `TestResolveOutlineLevels_SparseGuard`,
`TestNewLevelContext_OutlineBranch`, `TestOutlineFromInputs`),
`tokenizer_unit_test.go` (`TestChunkDocsToMaps_PreservesPDFPositions`),
`token_pdfpos_test.go`.
- **Batch 1** — `token_batch1_test.go`:
`TestSentenceDelimiterMatchesBangAndQuestion`,
`TestMergeByTokenSizeFromJSON_OverlapStripsTags`,
`TestMergeByTokenSizeFromJSON_EmptyPrevKeepsChunk`,
`TestTakeFromEndRespectsTokenCount`,
`TestTakeFromStartRespectsTokenCount`.
- **Batch 2** — `qa_batch2_test.go`:
`TestQAChunker_DefaultLangIsChinese`,
`TestRmQAPrefixStripsMultipleSeparators`, `TestQAChunker_SetsTopInt`,
`TestQAChunker_CarriesImageAndPositions`. Existing `qa_test.go`
expectations were updated to the corrected language default / separator
behavior.
- `pipeline_real_integration_test.go`
(`TestPipelineExecutor_Run_RealCanvasDSL_UsesGeneralPipeline`,
`TestPipelineExecutor_Run_RealPDF_ProducesIndexedChunks`,
`TestRunPipeline_RealPipelineOutput_ProducesIndexFields`) now runs
without any external service.
- `bash build.sh --test ./internal/ingestion/...` passes.
- No files deleted.
## Summary
Three groups of changes across the Go ingestion pipeline:
### 1. DOCX parsing improvements
- **docx_parser.go**: Enhanced DOCX parsing with better structure
extraction and media handling
- **docx_parser_cgo_test.go**, **docx_parser_test.go**: Companion tests
- **office_parsers_no_cgo.go**: Stub sync for non-CGO builds
### 2. Email (.eml) parsing: base64 Content-Transfer-Encoding decoding
- **email_parser.go** (`decodeCTE`): Added Content-Transfer-Encoding
decoding for base64 and quoted-printable. Go's `mime/multipart.Reader`
does not decode Content-Transfer-Encoding automatically, so attachments
with `Content-Transfer-Encoding: base64` remained base64-encoded in the
output. The new `decodeCTE` helper is called after reading each
multipart part's raw bytes in `readMailBody`, mirroring Python's
`part.get_payload(decode=True)`.
- **email_parser_test.go**: Two new tests — simple base64 attachment and
nested multipart/alternative with base64 attachment.
### 3. Extractor LLM driver fix + ModelDriver consolidation
- **extractor.go**: Fixed a bug where the Extractor component used
`ModelFactory.CreateModelDriver()`, which creates bare model instances
without API keys or provider configuration. Switched to
`models.GetPreconfiguredDriver()` which resolves the actual
pre-configured driver from `ProviderManager`, matching the codepath used
by `llm.go`. This fixes auto keyword/question extraction in DSL
pipelines that require LLM calls.
- **get_driver.go** (new): Extracted shared `GetPreconfiguredDriver()`
from `llm.go:newChatModelDriver()` so both `llm.go` and `extractor.go`
use the same codepath.
- **get_driver_test.go** (new): Tests for the shared driver resolution.
- **llm.go**: Replaced inline driver resolution with
`models.GetPreconfiguredDriver()`.
### 4. Chunker fixes and observability
- **group.go** (`extractLineRecords`): Fixed to also read `markdown` and
`html` payload keys — previously it only read `text`/`content`, causing
GroupTitleChunker to silently return empty results for markdown-format
parser output.
- **common.go** (`compileDelimPattern`): Aligned with Python's
`_compile_delimiter_pattern` — only backtick-wrapped delimiters produce
an active regex pattern; plain delimiters are not compiled into the
split regex.
- **token.go** (`applyChildrenDelim`): Set `DocType` and `CKType` to
`"text"` on created ChunkDocs so the token-size merge path correctly
identifies and merges text segments.
- **parser.go**, **extractor.go**, **tokenizer.go**, **group.go**,
**hierarchy.go**: Added debug-level logging for pipeline diagnostics.
- **parser_dispatch_test.go**, **group_test.go**: New tests.
## Verification
- All Go tests pass: `bash build.sh --test ./internal/parser/parser/...`
and `bash build.sh --test ./internal/ingestion/component/...`
- Build succeeds: `bash build.sh --go`
Ports two Python fixes to Go: the variable_ref_patt underscore/colon fix
(#16792) and the TokenChunker upstream-chunks fix (#16825). Keeps Go
behavior aligned with upstream Python.
Two refactors on the Go port (agent-go-port):
- Remove the dead per-tenant canvas runtime selector (write-only Redis
scaffolding with no runtime callers) and its dependent metrics/admin
code.
- Move the tokenizer embedding-model id from the shared ingestion
globals into a Tokenizer-scoped setup, and wire the production embedder
resolver in the ingestion task package.
32 files changed, 861 insertions, 1228 deletions.
feat(ingestion): mirror Go pipeline progress into the document table;
harden resume guards
- pipeline: bind the owning document via WithDocumentID; after each
TrackProgress event aggregate ingestion_task_log progress and mirror
progress/run/progress_msg back into the document table, so GET
/api/v1/datasets/{dataset_id}/documents reflects live Go pipeline
progress without a bespoke endpoint.
- canvas: extend the S3 resume guard to reject legacy no-op nodes (e.g.
ExitLoop) so component_total equals the count of progress-reporting
components and the aggregate percent can reach 100%.
- runtime/canvas: route progress through TrackProgress; add interrupt
test coverage (r3_interrupt_test.go).
- dao/entity: add IngestionTask.DocumentID column and AggregateProgress
support used by the mirror; IngestionTaskLog keeps a Checkpoint column
alongside the progress fields.
feat(deepdoc): cache DocAnalyzer inference results in Redis (1h TTL)
- Redis-backed DocAnalyzerCache decorator over inference.Client; cache
key = "ddoc:cache:<method>:" + sha256 of the JPEG-encoded image bytes
(deterministic).
- TTL = 1h; hits skip the inner HTTP call and return cached JSON; inner
errors are not cached.
refactor(deepdoc): align figure cropping with Python cropout + bounded
page caches
- CropSectionByDLA mirrors Python cropout: best-overlap DLA
figure/equation region, fallback to section bbox per page, vertical
concat on gray background.
- sliding-window page-image cache bounds peak memory to the recent
window instead of the whole PDF.
- rename DLADebug -> DLARegions across parser/chunker/tests.
refactor(parser): drop lib_type selector; align NewXxxParser with
NewPDFParser
- remove config["lib_type"] lookup and the libType param/field/switch
from all nine constructors; surface the CGO-required error at
ParseWithResult time instead of construction time; drop resolveLibType,
its test, and the four lib_type constants.
feat(utility): add a reusable workerpool for bounded concurrent
execution
- internal/utility/workerpool.go (+ tests).
refactor: translate Chinese prose comments to English in non-harness Go
files.
chore: upgrade github.com/cloudwego/eino from v0.9.9 to v0.9.12.
