Merge pull request #15903 from Hisma/marker-api-update

feat: Add configurable API URL (for self-hosting) and additional_config parameter for Datalab Marker API
This commit is contained in:
Tim Jaeryang Baek
2025-08-04 15:21:03 +04:00
committed by GitHub
6 changed files with 215 additions and 98 deletions
@@ -15,24 +15,28 @@ class DatalabMarkerLoader:
self,
file_path: str,
api_key: str,
langs: Optional[str] = None,
api_base_url: str,
additional_config: Optional[str] = None,
use_llm: bool = False,
skip_cache: bool = False,
force_ocr: bool = False,
paginate: bool = False,
strip_existing_ocr: bool = False,
disable_image_extraction: bool = False,
format_lines: bool = False,
output_format: str = None,
):
self.file_path = file_path
self.api_key = api_key
self.langs = langs
self.api_base_url = api_base_url
self.additional_config = additional_config
self.use_llm = use_llm
self.skip_cache = skip_cache
self.force_ocr = force_ocr
self.paginate = paginate
self.strip_existing_ocr = strip_existing_ocr
self.disable_image_extraction = disable_image_extraction
self.format_lines = format_lines
self.output_format = output_format
def _get_mime_type(self, filename: str) -> str:
@@ -60,7 +64,7 @@ class DatalabMarkerLoader:
return mime_map.get(ext, "application/octet-stream")
def check_marker_request_status(self, request_id: str) -> dict:
url = f"https://www.datalab.to/api/v1/marker/{request_id}"
url = f"{self.api_base_url}/{request_id}"
headers = {"X-Api-Key": self.api_key}
try:
response = requests.get(url, headers=headers)
@@ -81,22 +85,25 @@ class DatalabMarkerLoader:
)
def load(self) -> List[Document]:
url = "https://www.datalab.to/api/v1/marker"
url = self.api_base_url
filename = os.path.basename(self.file_path)
mime_type = self._get_mime_type(filename)
headers = {"X-Api-Key": self.api_key}
form_data = {
"langs": self.langs,
"use_llm": str(self.use_llm).lower(),
"skip_cache": str(self.skip_cache).lower(),
"force_ocr": str(self.force_ocr).lower(),
"paginate": str(self.paginate).lower(),
"strip_existing_ocr": str(self.strip_existing_ocr).lower(),
"disable_image_extraction": str(self.disable_image_extraction).lower(),
"format_lines": str(self.format_lines).lower(),
"output_format": self.output_format,
}
if self.additional_config and self.additional_config.strip():
form_data["additional_config"] = self.additional_config
log.info(
f"Datalab Marker POST request parameters: {{'filename': '{filename}', 'mime_type': '{mime_type}', **{form_data}}}"
)
@@ -133,74 +140,92 @@ class DatalabMarkerLoader:
check_url = result.get("request_check_url")
request_id = result.get("request_id")
if not check_url:
raise HTTPException(
status.HTTP_502_BAD_GATEWAY, detail="No request_check_url returned."
)
for _ in range(300): # Up to 10 minutes
time.sleep(2)
try:
poll_response = requests.get(check_url, headers=headers)
poll_response.raise_for_status()
poll_result = poll_response.json()
except (requests.HTTPError, ValueError) as e:
raw_body = poll_response.text
log.error(f"Polling error: {e}, response body: {raw_body}")
raise HTTPException(
status.HTTP_502_BAD_GATEWAY, detail=f"Polling failed: {e}"
)
status_val = poll_result.get("status")
success_val = poll_result.get("success")
if status_val == "complete":
summary = {
k: poll_result.get(k)
for k in (
"status",
"output_format",
"success",
"error",
"page_count",
"total_cost",
# Check if this is a direct response (self-hosted) or polling response (DataLab)
if check_url:
# DataLab polling pattern
for _ in range(300): # Up to 10 minutes
time.sleep(2)
try:
poll_response = requests.get(check_url, headers=headers)
poll_response.raise_for_status()
poll_result = poll_response.json()
except (requests.HTTPError, ValueError) as e:
raw_body = poll_response.text
log.error(f"Polling error: {e}, response body: {raw_body}")
raise HTTPException(
status.HTTP_502_BAD_GATEWAY, detail=f"Polling failed: {e}"
)
}
log.info(
f"Marker processing completed successfully: {json.dumps(summary, indent=2)}"
)
break
if status_val == "failed" or success_val is False:
log.error(
f"Marker poll failed full response: {json.dumps(poll_result, indent=2)}"
)
error_msg = (
poll_result.get("error")
or "Marker returned failure without error message"
status_val = poll_result.get("status")
success_val = poll_result.get("success")
if status_val == "complete":
summary = {
k: poll_result.get(k)
for k in (
"status",
"output_format",
"success",
"error",
"page_count",
"total_cost",
)
}
log.info(
f"Marker processing completed successfully: {json.dumps(summary, indent=2)}"
)
break
if status_val == "failed" or success_val is False:
log.error(
f"Marker poll failed full response: {json.dumps(poll_result, indent=2)}"
)
error_msg = (
poll_result.get("error")
or "Marker returned failure without error message"
)
raise HTTPException(
status.HTTP_400_BAD_REQUEST,
detail=f"Marker processing failed: {error_msg}",
)
else:
raise HTTPException(
status.HTTP_504_GATEWAY_TIMEOUT,
detail="Marker processing timed out",
)
if not poll_result.get("success", False):
error_msg = poll_result.get("error") or "Unknown processing error"
raise HTTPException(
status.HTTP_400_BAD_REQUEST,
detail=f"Marker processing failed: {error_msg}",
detail=f"Final processing failed: {error_msg}",
)
# DataLab format - content in format-specific fields
content_key = self.output_format.lower()
raw_content = poll_result.get(content_key)
final_result = poll_result
else:
raise HTTPException(
status.HTTP_504_GATEWAY_TIMEOUT, detail="Marker processing timed out"
)
# Self-hosted direct response - content in "output" field
if "output" in result:
log.info("Self-hosted Marker returned direct response without polling")
raw_content = result.get("output")
final_result = result
else:
available_fields = (
list(result.keys())
if isinstance(result, dict)
else "non-dict response"
)
raise HTTPException(
status.HTTP_502_BAD_GATEWAY,
detail=f"Custom Marker endpoint returned success but no 'output' field found. Available fields: {available_fields}. Expected either 'request_check_url' for polling or 'output' field for direct response.",
)
if not poll_result.get("success", False):
error_msg = poll_result.get("error") or "Unknown processing error"
raise HTTPException(
status.HTTP_400_BAD_REQUEST,
detail=f"Final processing failed: {error_msg}",
)
content_key = self.output_format.lower()
raw_content = poll_result.get(content_key)
if content_key == "json":
if self.output_format.lower() == "json":
full_text = json.dumps(raw_content, indent=2)
elif content_key in {"markdown", "html"}:
elif self.output_format.lower() in {"markdown", "html"}:
full_text = str(raw_content).strip()
else:
raise HTTPException(
@@ -211,14 +236,14 @@ class DatalabMarkerLoader:
if not full_text:
raise HTTPException(
status.HTTP_400_BAD_REQUEST,
detail="Datalab Marker returned empty content",
detail="Marker returned empty content",
)
marker_output_dir = os.path.join("/app/backend/data/uploads", "marker_output")
os.makedirs(marker_output_dir, exist_ok=True)
file_ext_map = {"markdown": "md", "json": "json", "html": "html"}
file_ext = file_ext_map.get(content_key, "txt")
file_ext = file_ext_map.get(self.output_format.lower(), "txt")
output_filename = f"{os.path.splitext(filename)[0]}.{file_ext}"
output_path = os.path.join(marker_output_dir, output_filename)
@@ -231,13 +256,13 @@ class DatalabMarkerLoader:
metadata = {
"source": filename,
"output_format": poll_result.get("output_format", self.output_format),
"page_count": poll_result.get("page_count", 0),
"output_format": final_result.get("output_format", self.output_format),
"page_count": final_result.get("page_count", 0),
"processed_with_llm": self.use_llm,
"request_id": request_id or "",
}
images = poll_result.get("images", {})
images = final_result.get("images", {})
if images:
metadata["image_count"] = len(images)
metadata["images"] = json.dumps(list(images.keys()))
+7 -1
View File
@@ -281,10 +281,15 @@ class Loader:
"tiff",
]
):
api_base_url = self.kwargs.get("DATALAB_MARKER_API_BASE_URL", "")
if not api_base_url or api_base_url.strip() == "":
api_base_url = "https://www.datalab.to/api/v1/marker"
loader = DatalabMarkerLoader(
file_path=file_path,
api_key=self.kwargs["DATALAB_MARKER_API_KEY"],
langs=self.kwargs.get("DATALAB_MARKER_LANGS"),
api_base_url=api_base_url,
additional_config=self.kwargs.get("DATALAB_MARKER_ADDITIONAL_CONFIG"),
use_llm=self.kwargs.get("DATALAB_MARKER_USE_LLM", False),
skip_cache=self.kwargs.get("DATALAB_MARKER_SKIP_CACHE", False),
force_ocr=self.kwargs.get("DATALAB_MARKER_FORCE_OCR", False),
@@ -295,6 +300,7 @@ class Loader:
disable_image_extraction=self.kwargs.get(
"DATALAB_MARKER_DISABLE_IMAGE_EXTRACTION", False
),
format_lines=self.kwargs.get("DATALAB_MARKER_FORMAT_LINES", False),
output_format=self.kwargs.get(
"DATALAB_MARKER_OUTPUT_FORMAT", "markdown"
),