Files
ragflow/internal/parser/parser/markdown_parser.go
Zhichang Yu 12787996d1 feat(agent): Go ingestion pipeline progress mirroring and DeepDOC parser hardening (#16795)
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.
2026-07-10 10:36:10 +08:00

151 lines
4.5 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, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
package parser
import (
"bytes"
"strings"
"github.com/gomarkdown/markdown/ast"
"github.com/gomarkdown/markdown/parser"
)
type MarkdownParser struct{}
func NewMarkdownParser() *MarkdownParser {
return &MarkdownParser{}
}
// ParseWithResult implements ParseResultProducer (plan §6.5) and
// returns a structured markdown payload that mirrors the Python
// parser's `output_format == "json"` shape. Each top-level block
// emits one item with `text` + `doc_type_kwd: "text"`. The legacy
// debug-print path has been removed; callers consume ParseResult directly.
func (p *MarkdownParser) ParseWithResult(filename string, data []byte) ParseResult {
doc := markdownNew().Parse(data)
var items []map[string]any
walkMarkdownBlocks(doc, &items)
if items == nil {
// No blocks emitted — surface a single empty item so the
// downstream chunker sees a non-nil JSON slice (mirrors the
// Python contract of always producing at least one item).
items = []map[string]any{{"text": "", "doc_type_kwd": "text"}}
}
return ParseResult{
OutputFormat: "json",
File: map[string]any{
"name": filename,
},
JSON: items,
}
}
func (p *MarkdownParser) String() string {
return "MarkdownParser"
}
// markdownNew is a thin constructor so the extension set is owned
// in one place (both Parse and ParseWithResult consume it).
func markdownNew() *parser.Parser {
extensions := parser.CommonExtensions | parser.AutoHeadingIDs | parser.NoEmptyLineBeforeBlock
return parser.NewWithExtensions(extensions)
}
// walkMarkdownBlocks emits one normalized item per top-level block.
// Headings (LeafBlock / H*) are emitted with the heading text so a
// downstream title chunker can find them; paragraphs (Paragraph)
// emit the leaf-node text. The walker is intentionally a
// best-effort pass — full TOC / outline handling lands with the
// deepdoc/parser port — but it satisfies the per-block "emit
// text+doc_type_kwd" contract enough for a Phase-1a migration
// test to verify wire-shape parity.
func walkMarkdownBlocks(doc ast.Node, out *[]map[string]any) {
for _, child := range doc.GetChildren() {
switch n := child.(type) {
case *ast.Heading:
*out = append(*out, map[string]any{
"text": headingText(n),
"doc_type_kwd": "text",
"ck_type": "heading",
})
case *ast.Paragraph:
*out = append(*out, map[string]any{
"text": leafText(n),
"doc_type_kwd": "text",
"ck_type": "text",
})
case *ast.List:
*out = append(*out, map[string]any{
"text": leafText(n),
"doc_type_kwd": "text",
"ck_type": "list",
})
case *ast.CodeBlock:
*out = append(*out, map[string]any{
"text": leafText(n),
"doc_type_kwd": "text",
"ck_type": "code",
})
default:
// Block types we don't yet normalize (HTML, tables,
// images, definitions) are best-effort: emit the leaf
// text without a ck_type so downstream components can
// still treat them as text chunks.
txt := leafText(n)
if strings.TrimSpace(txt) != "" {
*out = append(*out, map[string]any{
"text": txt,
"doc_type_kwd": "text",
})
}
}
}
}
// headingText returns the inline-text of a heading node by
// concatenating every Leaf / Text child. Empty headings emit "".
func headingText(h *ast.Heading) string {
var buf bytes.Buffer
for _, c := range h.GetChildren() {
buf.WriteString(leafText(c))
}
return strings.TrimSpace(buf.String())
}
// leafText mirrors gomarkdown's leaf walker: walks every descendant
// leaf (Text or Inline content) and returns the concatenated UTF-8.
// Non-text containers that have no leaf descendants return "".
func leafText(n ast.Node) string {
var buf bytes.Buffer
walkLeaf(n, &buf)
return strings.TrimSpace(buf.String())
}
func walkLeaf(n ast.Node, buf *bytes.Buffer) {
switch t := n.(type) {
case *ast.Text:
buf.Write(t.Literal)
case *ast.Code:
buf.Write(t.Literal)
default:
for _, c := range n.GetChildren() {
walkLeaf(c, buf)
}
}
}