mirror of
https://github.com/infiniflow/ragflow.git
synced 2026-08-02 22:07:31 +08:00
## Summary
Adds a side-effect-free DataFlow canvas **debug (dry-run) mode** plus a
**debug run log with a "View result" panel**, so a canvas can be
executed synchronously and inspected end-to-end (per-component progress
and parsed chunks) without persisting anything.
### Dry-run execution (inline parsed chunks)
- `task/debug.go`: `NewDebugTaskContext` builds an in-memory
`TaskContext` with **`KB.ID == ""` — the single debug signal used across
the ingestion pipeline**. A canvas debug run has no knowledgebase, and
production ingestion always supplies one, so `kb_id == ""` occurs ONLY
in debug mode. Components gate their own side effects on this signal
without any dedicated debug vocabulary (the former `CANVAS_DEBUG_DOC_ID`
marker constant is removed).
- `task/pipeline_executor.go`: `validateTaskContext` no longer requires
a KB when `KB.ID == ""` (debug); debug runs return `collectDebugOutput`
(chunks) instead of a no-op; uploaded bytes are delivered as
`inputs['binary']` for doc-less runs; `injectDebugPageCap` caps the
parser to the first pages for a fast preview via the production
`override_params` channel (Parser cpnID + family).
- `component/tokenizer.go`: `shouldHaveEmbedding` skips embedding when
`kb_id == ""` — the embedder is configured on the knowledgebase, so a
debug run has nothing to resolve against and stays side-effect free.
- `chunker/register.go`: chunk images are uploaded to MinIO only when a
KB is present (persist run). **In debug mode the raw image bytes are
intentionally dropped (`delete(ck, "image")`)** — the debug preview does
not render chunk images, and dropping the bytes keeps them out of memory
and out of the Redis-stored debug log. This is a deliberate trade-off,
not an oversight.
- `component/file.go`: pass through in-memory binary bytes, skipping
`doc_id` -> storage resolution.
- `handler/agent.go` + `agent_webhook.go`: detect `dataflow_canvas` and
run a sync debug returning chunks inline on the existing
chat/completions endpoint; reject DataFlow canvases from webhooks (fixes
the previously dead `== "DataFlow"` check; mirrors Python
`agent_api.py`).
- `parser_dispatch.go`: export `ParserFileFamily` for the executor's
page-cap injection.
### Debug run log + "View result"
Mirrors Python's debug-log contract so the front-end can replay each
component's progress and parsed output:
- `task/debug_log_sink.go`: a `DebugLogSink` records every component's
lifecycle into a `[{component_id, trace}]` array (each trace entry
carries `message`, `progress`, `timestamp`, `elapsed_time`). `Flush`
appends a terminal `END` marker whose first trace message is non-empty
so the front-end detects completion. On failure the END marker is
prefixed `[ERROR]` yet still carries the run, so the failure timeline
renders instead of being stuck empty. Timestamps and `elapsed_time` are
in seconds (matching the rest of the app).
- `task/debug_result_dsl.go`: `BuildDebugResultDSL` builds the `dsl` the
END marker carries — the Go analogue of Python's `Graph.__str__` +
END-marker `dsl` in `rag/flow/pipeline.py`. It combines the static DSL
structure (component_name / downstream / params / graph.nodes) with the
run output map (`output["state"][<id>]`) to emit, per component,
`obj.params.outputs[<format>].value` (chunks / text / json / html /
markdown) — the exact keys the front-end `dataflow-result` page reads to
render each step's parsed chunks. Raw embedding vectors (including the
dimension-scoped `q_<dim>_vec` keys) are stripped so the stored log
stays Python-scale.
- `task/pipeline_executor.go`: after the run, attach the built `dsl` to
the END marker via the `ResultSink` capability.
- `handler/agent.go`: `runCanvasPipelineDebug` generates a stable
`message_id` up-front and always flushes the log (success or failure);
`respondWithDebugResult` returns `message_id` in **both** the success
and the error envelope so the front-end can poll the log. The debug-log
endpoint `GET /agents/:id/logs/:message_id` serves the array.
- `web/src/pages/agent/hooks/use-run-dataflow.ts`: on a run failure,
also surface `message_id` via `setMessageId` so the log sheet renders
the failure timeline (the `[ERROR]` END marker is already written).
Guarded by `if (msgId)`, so it is a safe no-op when the back-end does
not return an id.
## Behavioral notes
- Debug parses only the first pages (`debugPageCapPages`) for a fast
preview; an explicit `pages` cap already present in the ParserConfig is
respected.
- Debug mode does not keep chunk images (see above) and does not compute
embeddings — it exercises parse + chunk only.
## Test plan
- Go: `debug_test.go`, `debug_log_sink_test.go` (trace pairing, END
marker, `[ERROR]` prefix, fractional-second timestamp/elapsed_time, size
caps, and a real-pipeline test asserting the END-marker `dsl` carries
non-empty per-component `params.outputs` with chunks),
`debug_result_dsl_test.go` (flat and real nested `output["state"]`
shapes, vector stripping, format priority),
`debug_pages_integration_test.go`, `pipeline_executor_persist_test.go`,
`handler/agent_pipeline_debug_test.go`, `handler/agent_logs_test.go`
(incl. `TestRunCanvasPipelineDebug_ErrorStillExposesMessageID` /
`TestRespondWithDebugResult_ErrorCarriesMessageID` locking `message_id`
on failure), plus updates to `agent_test.go` / `agent_webhook_test.go` /
`chunker/image_upload_test.go` / `tokenizer*.go`.
- `go build ./...` and `./build.sh --test` for affected packages.
🤖 Generated with [CodeBuddy Code](https://cnb.cool/codebuddy)
---------
Co-authored-by: CodeBuddy Code <noreply@tencent.com>
324 lines
11 KiB
Go
324 lines
11 KiB
Go
//
|
|
// Copyright 2026 The InfiniFlow Authors. All Rights Reserved.
|
|
//
|
|
// Licensed under the Apache License, Version 2.0 (the "License");
|
|
// you may not use this file except in compliance with the License.
|
|
// You may obtain a copy of the License at
|
|
//
|
|
// http://www.apache.org/licenses/LICENSE-2.0
|
|
//
|
|
// Unless required by applicable law or agreed to in writing, software
|
|
// distributed under the License is distributed on an "AS IS" BASIS,
|
|
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, or express or implied.
|
|
// See the License for the specific language governing permissions and
|
|
// limitations under the License.
|
|
//
|
|
|
|
package chunker
|
|
|
|
import (
|
|
"context"
|
|
"encoding/base64"
|
|
"fmt"
|
|
"sync"
|
|
"sync/atomic"
|
|
"testing"
|
|
"time"
|
|
|
|
"ragflow/internal/common"
|
|
)
|
|
|
|
const (
|
|
testKBID = "kb-1"
|
|
testDocID = "doc-1"
|
|
)
|
|
|
|
// pngBase64 is a tiny 1x1 PNG payload (raw base64, no data-URL prefix).
|
|
const pngBase64 = "iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg=="
|
|
|
|
// TestUploadOneImage_UploadsBytes pins the pure upload contract: it receives
|
|
// already-decoded bytes (not the raw chunk image field) plus the chunkID, and
|
|
// returns the img_id — it does NOT touch any chunk map.
|
|
func TestUploadOneImage_UploadsBytes(t *testing.T) {
|
|
var gotBucket, gotKey string
|
|
var gotData []byte
|
|
up := func(_ context.Context, kbID, chunkID string, data []byte) (string, error) {
|
|
gotBucket, gotKey, gotData = kbID, chunkID, data
|
|
return fmt.Sprintf("%s-%s", kbID, chunkID), nil
|
|
}
|
|
chunkID := common.ChunkID(testDocID, "caption")
|
|
data, _ := base64.StdEncoding.DecodeString(pngBase64)
|
|
|
|
imgID, err := uploadOneImage(context.Background(), up, testKBID, chunkID, data)
|
|
if err != nil {
|
|
t.Fatalf("uploadOneImage: %v", err)
|
|
}
|
|
if imgID != testKBID+"-"+chunkID {
|
|
t.Errorf("img_id = %q, want %q", imgID, testKBID+"-"+chunkID)
|
|
}
|
|
if gotBucket != testKBID || gotKey != chunkID {
|
|
t.Errorf("upload target = (%q,%q), want (%q,%q)", gotBucket, gotKey, testKBID, chunkID)
|
|
}
|
|
if string(gotData) != string(data) {
|
|
t.Errorf("uploaded bytes mismatch")
|
|
}
|
|
}
|
|
|
|
// TestUploadChunkImages_WritesImgIDAndDropsImage pins the caller-side contract:
|
|
// uploadChunkImages reads ck["id"] (already written by the decorator prior to
|
|
// calling upload), decodes the chunk's image, uploads via uploadOneImage, then
|
|
// writes ck["img_id"] and deletes the raw image field.
|
|
func TestUploadChunkImages_WritesImgIDAndDropsImage(t *testing.T) {
|
|
var gotKey string
|
|
up := func(_ context.Context, kbID, chunkID string, _ []byte) (string, error) {
|
|
gotKey = chunkID
|
|
return kbID + "-" + chunkID, nil
|
|
}
|
|
chunkID := common.ChunkID(testDocID, "caption")
|
|
ck := map[string]any{"id": chunkID, "content_with_weight": "caption", "image": "data:image/png;base64," + pngBase64}
|
|
|
|
if err := uploadChunkImages(context.Background(), []map[string]any{ck}, up, testKBID); err != nil {
|
|
t.Fatalf("uploadChunkImages: %v", err)
|
|
}
|
|
wantImgID := testKBID + "-" + chunkID
|
|
if ck["img_id"] != wantImgID {
|
|
t.Errorf("img_id = %v, want %q", ck["img_id"], wantImgID)
|
|
}
|
|
if ck["id"] != chunkID {
|
|
t.Errorf("id = %v, want %q", ck["id"], chunkID)
|
|
}
|
|
if _, ok := ck["image"]; ok {
|
|
t.Errorf("image field not deleted: %v", ck["image"])
|
|
}
|
|
if gotKey != chunkID {
|
|
t.Errorf("upload key = %q, want %q", gotKey, chunkID)
|
|
}
|
|
}
|
|
|
|
// TestUploadChunkImages_SkipsWhenImgIDPresent mirrors Python's
|
|
// `if d.get("img_id"): return` bypass — a chunk that already carries an
|
|
// img_id is left untouched and no upload happens.
|
|
func TestUploadChunkImages_SkipsWhenImgIDPresent(t *testing.T) {
|
|
called := false
|
|
up := func(_ context.Context, kbID, chunkID string, _ []byte) (string, error) {
|
|
called = true
|
|
return "should-not-be-used", nil
|
|
}
|
|
chunkID := common.ChunkID(testDocID, "x")
|
|
ck := map[string]any{"id": chunkID, "content_with_weight": "x", "img_id": "preset-img", "image": "data:image/png;base64," + pngBase64}
|
|
if err := uploadChunkImages(context.Background(), []map[string]any{ck}, up, testKBID); err != nil {
|
|
t.Fatalf("uploadChunkImages: %v", err)
|
|
}
|
|
if called {
|
|
t.Error("uploader was called; expected skip when img_id preset")
|
|
}
|
|
if ck["img_id"] != "preset-img" {
|
|
t.Errorf("img_id = %v, want preserved %q", ck["img_id"], "preset-img")
|
|
}
|
|
if ck["id"] != chunkID {
|
|
t.Errorf("id = %v, want preserved %q", ck["id"], chunkID)
|
|
}
|
|
}
|
|
|
|
// TestUploadChunkImages_NoImage: a chunk with no image gets img_id="" and no
|
|
// upload (Python: img_id="" branch). The id is preserved.
|
|
func TestUploadChunkImages_NoImage(t *testing.T) {
|
|
called := false
|
|
up := func(_ context.Context, _, _ string, _ []byte) (string, error) {
|
|
called = true
|
|
return "", nil
|
|
}
|
|
chunkID := common.ChunkID(testDocID, "text only")
|
|
ck := map[string]any{"id": chunkID, "content_with_weight": "text only"}
|
|
if err := uploadChunkImages(context.Background(), []map[string]any{ck}, up, testKBID); err != nil {
|
|
t.Fatalf("uploadChunkImages: %v", err)
|
|
}
|
|
if called {
|
|
t.Error("uploader was called for a chunk without image")
|
|
}
|
|
if ck["img_id"] != "" {
|
|
t.Errorf("img_id = %v, want empty", ck["img_id"])
|
|
}
|
|
if ck["id"] != chunkID {
|
|
t.Errorf("id = %v, want preserved %q", ck["id"], chunkID)
|
|
}
|
|
}
|
|
|
|
// TestUploadChunkImage_ErrorOnMissingID verifies that uploadChunkImage errors
|
|
// when ck["id"] is absent — the caller (decorator) must always set id first.
|
|
func TestUploadChunkImage_ErrorOnMissingID(t *testing.T) {
|
|
up := func(_ context.Context, _, _ string, _ []byte) (string, error) {
|
|
return "should-not-be-called", nil
|
|
}
|
|
ck := map[string]any{"content_with_weight": "x", "image": "data:image/png;base64," + pngBase64}
|
|
err := uploadChunkImages(context.Background(), []map[string]any{ck}, up, testKBID)
|
|
if err == nil {
|
|
t.Fatal("expected error when chunk missing id")
|
|
}
|
|
}
|
|
|
|
// TestUploadOneImage_ConcurrencyLimit verifies the process-wide semaphore caps
|
|
// concurrent uploads at the configured limit.
|
|
func TestUploadOneImage_ConcurrencyLimit(t *testing.T) {
|
|
const limit = 3
|
|
setImageUploadConcurrencyForTest(t, limit)
|
|
|
|
var inflight, maxSeen int64
|
|
release := make(chan struct{})
|
|
up := func(_ context.Context, kbID, chunkID string, _ []byte) (string, error) {
|
|
cur := atomic.AddInt64(&inflight, 1)
|
|
for {
|
|
m := atomic.LoadInt64(&maxSeen)
|
|
if cur <= m || atomic.CompareAndSwapInt64(&maxSeen, m, cur) {
|
|
break
|
|
}
|
|
}
|
|
<-release
|
|
atomic.AddInt64(&inflight, -1)
|
|
return fmt.Sprintf("%s-%s", kbID, chunkID), nil
|
|
}
|
|
|
|
const n = 12
|
|
var wg sync.WaitGroup
|
|
for i := 0; i < n; i++ {
|
|
wg.Add(1)
|
|
go func(i int) {
|
|
defer wg.Done()
|
|
data, _ := base64.StdEncoding.DecodeString(pngBase64)
|
|
chunkID := common.ChunkID(testDocID, fmt.Sprintf("c%d", i))
|
|
_, _ = uploadOneImage(context.Background(), up, testKBID, chunkID, data)
|
|
}(i)
|
|
}
|
|
waitForInflight(t, &inflight, limit)
|
|
close(release)
|
|
wg.Wait()
|
|
|
|
if maxSeen > limit {
|
|
t.Errorf("max concurrent uploads = %d, want <= %d", maxSeen, limit)
|
|
}
|
|
if maxSeen < limit {
|
|
t.Errorf("max concurrent uploads = %d, expected to reach limit %d", maxSeen, limit)
|
|
}
|
|
}
|
|
|
|
// TestImageUploadDecorator_EndToEnd drives a real OneChunker through the
|
|
// registration decorator and asserts the produced chunk was uploaded (img_id
|
|
// and id set, image dropped). Uses the ChunkImageUploader override seam so no
|
|
// real storage backend is needed.
|
|
func TestImageUploadDecorator_EndToEnd(t *testing.T) {
|
|
var gotKB, gotKey string
|
|
prev := ChunkImageUploader
|
|
ChunkImageUploader = func(_ context.Context, kbID, chunkID string, _ []byte) (string, error) {
|
|
gotKB, gotKey = kbID, chunkID
|
|
return kbID + "-" + chunkID, nil
|
|
}
|
|
t.Cleanup(func() { ChunkImageUploader = prev })
|
|
|
|
comp, err := NewOneChunker(nil)
|
|
if err != nil {
|
|
t.Fatalf("NewOneChunker: %v", err)
|
|
}
|
|
decorated := &imageUploadDecorator{inner: comp}
|
|
|
|
inputs := map[string]any{
|
|
"name": "doc.pdf",
|
|
"kb_id": testKBID,
|
|
"doc_id": testDocID,
|
|
"chunks": []map[string]any{
|
|
{"content_with_weight": "a cropped figure", "image": "data:image/png;base64," + pngBase64},
|
|
},
|
|
}
|
|
out, err := decorated.Invoke(context.Background(), nil, inputs)
|
|
if err != nil {
|
|
t.Fatalf("decorated Invoke: %v", err)
|
|
}
|
|
chunks, ok := out["chunks"].([]map[string]any)
|
|
if !ok || len(chunks) == 0 {
|
|
t.Fatalf("no chunks in output")
|
|
}
|
|
ck := chunks[0]
|
|
wantChunkID := common.ChunkID(testDocID, "a cropped figure")
|
|
if ck["img_id"] != testKBID+"-"+wantChunkID {
|
|
t.Errorf("img_id = %v, want %q", ck["img_id"], testKBID+"-"+wantChunkID)
|
|
}
|
|
if ck["id"] != wantChunkID {
|
|
t.Errorf("id = %v, want chunk id %q written back", ck["id"], wantChunkID)
|
|
}
|
|
if _, ok := ck["image"]; ok {
|
|
t.Errorf("image not dropped: %v", ck["image"])
|
|
}
|
|
if gotKB != testKBID || gotKey != wantChunkID {
|
|
t.Errorf("upload target = (%q,%q), want (%q,%q)", gotKB, gotKey, testKBID, wantChunkID)
|
|
}
|
|
}
|
|
|
|
// setImageUploadConcurrencyForTest swaps the process-wide upload semaphore to
|
|
// `n` slots for the duration of the test, restoring the original after.
|
|
func setImageUploadConcurrencyForTest(t *testing.T, n int) {
|
|
t.Helper()
|
|
prev := imageUploadSem
|
|
imageUploadSem = make(chan struct{}, n)
|
|
t.Cleanup(func() { imageUploadSem = prev })
|
|
}
|
|
|
|
// waitForInflight blocks until the atomic counter reaches `target` or fails.
|
|
func waitForInflight(t *testing.T, counter *int64, target int64) {
|
|
t.Helper()
|
|
deadline := time.After(2 * time.Second)
|
|
for {
|
|
if atomic.LoadInt64(counter) >= target {
|
|
return
|
|
}
|
|
select {
|
|
case <-deadline:
|
|
t.Fatalf("timed out waiting for %d inflight uploads (got %d)", target, atomic.LoadInt64(counter))
|
|
case <-time.After(time.Millisecond):
|
|
}
|
|
}
|
|
}
|
|
|
|
// TestImageUploadDecorator_DebugSkipsUpload verifies that a debug/dry-run run
|
|
// (empty kb_id) does NOT upload the chunk image to storage, while still
|
|
// dropping the raw image bytes. This keeps the debug path free of MinIO side
|
|
// effects and preserves the current memory behaviour (raw bytes discarded, not
|
|
// held until a persist stage). An empty kb_id only occurs in canvas debug
|
|
// (dry-run) mode; production ingestion always supplies a KB.
|
|
func TestImageUploadDecorator_DebugSkipsUpload(t *testing.T) {
|
|
prev := ChunkImageUploader
|
|
ChunkImageUploader = func(_ context.Context, _, _ string, _ []byte) (string, error) {
|
|
t.Fatalf("ChunkImageUploader must not be called in a debug run")
|
|
return "", nil
|
|
}
|
|
t.Cleanup(func() { ChunkImageUploader = prev })
|
|
|
|
comp, err := NewOneChunker(nil)
|
|
if err != nil {
|
|
t.Fatalf("NewOneChunker: %v", err)
|
|
}
|
|
decorated := &imageUploadDecorator{inner: comp}
|
|
|
|
inputs := map[string]any{
|
|
"name": "doc.pdf",
|
|
"kb_id": "", // debug mode: no KB -> no upload
|
|
"doc_id": testDocID,
|
|
"chunks": []map[string]any{
|
|
{"content_with_weight": "a cropped figure", "image": "data:image/png;base64," + pngBase64},
|
|
},
|
|
}
|
|
out, err := decorated.Invoke(context.Background(), nil, inputs)
|
|
if err != nil {
|
|
t.Fatalf("decorated Invoke: %v", err)
|
|
}
|
|
chunks, ok := out["chunks"].([]map[string]any)
|
|
if !ok || len(chunks) == 0 {
|
|
t.Fatalf("expected chunks in output, got %#v", out)
|
|
}
|
|
ck := chunks[0]
|
|
if _, stillHas := ck["image"]; stillHas {
|
|
t.Errorf("raw image should be dropped in debug run, but ck[\"image\"] is still present")
|
|
}
|
|
if id, _ := ck["img_id"].(string); id != "" {
|
|
t.Errorf("img_id should not be set in debug run, got %q", id)
|
|
}
|
|
}
|