From 5db1b296fbea33ddf5bca78d9d69afeab403cfb3 Mon Sep 17 00:00:00 2001
From: Yufeng He <40085740+he-yufeng@users.noreply.github.com>
Date: Thu, 4 Jun 2026 13:42:58 +0800
Subject: [PATCH] fix: fall back from empty Docling native chunks (#15601)

## Summary
- keep the native Docling chunking path when it returns usable chunks
- fall back to the standard Docling response parser when a chunked
request gets HTTP 200 but returns no usable chunks
- add a regression test for older Docling servers that accept the
chunking request but return a standard conversion payload

## Why
Older external Docling servers can accept a request containing
`do_chunking: true` and still return the standard conversion response
shape. The current code treats any HTTP 200 from the chunked request as
a native chunk response, finds no chunk entries, and returns zero
sections without trying the standard response parser.

Fixes #15569.

## Validation
- `python -m pytest
test\\unit_test\\deepdoc\\parser\\test_docling_parser_remote.py -q`
- `python -m py_compile deepdoc\\parser\\docling_parser.py
test\\unit_test\\deepdoc\\parser\\test_docling_parser_remote.py`
- `python -m ruff check deepdoc\\parser\\docling_parser.py
test\\unit_test\\deepdoc\\parser\\test_docling_parser_remote.py`
- `git diff --check`
---
 deepdoc/parser/docling_parser.py              |  5 +-
 .../parser/test_docling_parser_remote.py      | 76 +++++++++++++++++++
 2 files changed, 80 insertions(+), 1 deletion(-)
 create mode 100644 test/unit_test/deepdoc/parser/test_docling_parser_remote.py

diff --git a/deepdoc/parser/docling_parser.py b/deepdoc/parser/docling_parser.py
index 948a7acb0c..097e9c9045 100644
--- a/deepdoc/parser/docling_parser.py
+++ b/deepdoc/parser/docling_parser.py
@@ -477,7 +477,10 @@ class DoclingParser(RAGFlowPdfParser):
                     
             if callback:
                 callback(0.95, f"[Docling] Native chunks received: {len(sections)}")
-            return sections, tables
+            if sections:
+                return sections, tables
+
+            self.logger.warning("[Docling] Native chunking returned no usable chunks; trying standard response parsing.")
 
         # --- FALLBACK: Standard RAGFlow parsing for older docling servers ---
         docs = self._extract_remote_document_entries(response_json)
diff --git a/test/unit_test/deepdoc/parser/test_docling_parser_remote.py b/test/unit_test/deepdoc/parser/test_docling_parser_remote.py
new file mode 100644
index 0000000000..e3410c65e2
--- /dev/null
+++ b/test/unit_test/deepdoc/parser/test_docling_parser_remote.py
@@ -0,0 +1,76 @@
+from __future__ import annotations
+
+import importlib.util
+import sys
+import types
+from pathlib import Path
+
+import pytest
+
+
+ROOT = Path(__file__).resolve().parents[4]
+
+
+class _Response:
+    status_code = 200
+    text = ""
+
+    def __init__(self, payload):
+        self._payload = payload
+
+    def json(self):
+        return self._payload
+
+
+def _load_docling_parser(monkeypatch):
+    common_pkg = types.ModuleType("common")
+    constants_mod = types.ModuleType("common.constants")
+    constants_mod.MAXIMUM_PAGE_NUMBER = 1000
+
+    deepdoc_pkg = types.ModuleType("deepdoc")
+    parser_pkg = types.ModuleType("deepdoc.parser")
+    parser_pkg.__path__ = []
+    utils_mod = types.ModuleType("deepdoc.parser.utils")
+    utils_mod.extract_pdf_outlines = lambda _source: []
+
+    pil_pkg = types.ModuleType("PIL")
+    image_mod = types.ModuleType("PIL.Image")
+    image_mod.Image = object
+    pil_pkg.Image = image_mod
+
+    monkeypatch.setitem(sys.modules, "common", common_pkg)
+    monkeypatch.setitem(sys.modules, "common.constants", constants_mod)
+    monkeypatch.setitem(sys.modules, "deepdoc", deepdoc_pkg)
+    monkeypatch.setitem(sys.modules, "deepdoc.parser", parser_pkg)
+    monkeypatch.setitem(sys.modules, "deepdoc.parser.utils", utils_mod)
+    monkeypatch.setitem(sys.modules, "pdfplumber", types.ModuleType("pdfplumber"))
+    monkeypatch.setitem(sys.modules, "PIL", pil_pkg)
+    monkeypatch.setitem(sys.modules, "PIL.Image", image_mod)
+
+    spec = importlib.util.spec_from_file_location(
+        "_docling_parser_under_test",
+        ROOT / "deepdoc" / "parser" / "docling_parser.py",
+    )
+    module = importlib.util.module_from_spec(spec)
+    monkeypatch.setitem(sys.modules, spec.name, module)
+    spec.loader.exec_module(module)
+    return module
+
+
+@pytest.mark.p2
+def test_remote_chunked_200_standard_payload_falls_back(monkeypatch):
+    module = _load_docling_parser(monkeypatch)
+    calls = []
+
+    def fake_post(_url, json, timeout):
+        calls.append((json, timeout))
+        return _Response({"document": {"md_content": "# Parsed\n\nbody"}})
+
+    monkeypatch.setattr(module.requests, "post", fake_post)
+
+    parser = module.DoclingParser(docling_server_url="http://docling.local")
+    sections, tables = parser._parse_pdf_remote("sample.pdf", binary=b"%PDF", parse_method="raw")
+
+    assert sections == [("# Parsed\n\nbody", "")]
+    assert tables == []
+    assert calls[0][0]["options"]["do_chunking"] is True