docling-serve 0.3.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- docling_serve-0.3.0/LICENSE +21 -0
- docling_serve-0.3.0/PKG-INFO +483 -0
- docling_serve-0.3.0/README.md +442 -0
- docling_serve-0.3.0/docling_serve/__init__.py +0 -0
- docling_serve-0.3.0/docling_serve/__main__.py +302 -0
- docling_serve-0.3.0/docling_serve/app.py +230 -0
- docling_serve-0.3.0/docling_serve/docling_conversion.py +431 -0
- docling_serve-0.3.0/docling_serve/gradio_ui.py +635 -0
- docling_serve-0.3.0/docling_serve/helper_functions.py +62 -0
- docling_serve-0.3.0/docling_serve/py.typed +0 -0
- docling_serve-0.3.0/docling_serve/response_preparation.py +248 -0
- docling_serve-0.3.0/docling_serve/settings.py +33 -0
- docling_serve-0.3.0/docling_serve.egg-info/PKG-INFO +483 -0
- docling_serve-0.3.0/docling_serve.egg-info/SOURCES.txt +22 -0
- docling_serve-0.3.0/docling_serve.egg-info/dependency_links.txt +1 -0
- docling_serve-0.3.0/docling_serve.egg-info/entry_points.txt +2 -0
- docling_serve-0.3.0/docling_serve.egg-info/requires.txt +28 -0
- docling_serve-0.3.0/docling_serve.egg-info/top_level.txt +1 -0
- docling_serve-0.3.0/pyproject.toml +180 -0
- docling_serve-0.3.0/setup.cfg +4 -0
- docling_serve-0.3.0/tests/test_1-file-all-outputs.py +129 -0
- docling_serve-0.3.0/tests/test_1-url-all-outputs.py +123 -0
- docling_serve-0.3.0/tests/test_2-files-all-outputs.py +74 -0
- docling_serve-0.3.0/tests/test_2-urls-all-outputs.py +67 -0
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
MIT License
|
|
2
|
+
|
|
3
|
+
Copyright (c) 2024 International Business Machines
|
|
4
|
+
|
|
5
|
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
6
|
+
of this software and associated documentation files (the "Software"), to deal
|
|
7
|
+
in the Software without restriction, including without limitation the rights
|
|
8
|
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
9
|
+
copies of the Software, and to permit persons to whom the Software is
|
|
10
|
+
furnished to do so, subject to the following conditions:
|
|
11
|
+
|
|
12
|
+
The above copyright notice and this permission notice shall be included in all
|
|
13
|
+
copies or substantial portions of the Software.
|
|
14
|
+
|
|
15
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
16
|
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
17
|
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
18
|
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
19
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
20
|
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
21
|
+
SOFTWARE.
|
|
@@ -0,0 +1,483 @@
|
|
|
1
|
+
Metadata-Version: 2.2
|
|
2
|
+
Name: docling-serve
|
|
3
|
+
Version: 0.3.0
|
|
4
|
+
Summary: Running Docling as a service
|
|
5
|
+
Author-email: Michele Dolfi <dol@zurich.ibm.com>, Guillaume Moutier <gmoutier@redhat.com>, Anil Vishnoi <avishnoi@redhat.com>, Panos Vagenas <pva@zurich.ibm.com>, Panos Vagenas <pva@zurich.ibm.com>, Christoph Auer <cau@zurich.ibm.com>, Peter Staar <taa@zurich.ibm.com>
|
|
6
|
+
Maintainer-email: Michele Dolfi <dol@zurich.ibm.com>, Anil Vishnoi <avishnoi@redhat.com>, Panos Vagenas <pva@zurich.ibm.com>, Christoph Auer <cau@zurich.ibm.com>, Peter Staar <taa@zurich.ibm.com>
|
|
7
|
+
License: MIT
|
|
8
|
+
Project-URL: Homepage, https://github.com/DS4SD/docling-serve
|
|
9
|
+
Project-URL: Repository, https://github.com/DS4SD/docling-serve
|
|
10
|
+
Project-URL: Issues, https://github.com/DS4SD/docling-serve/issues
|
|
11
|
+
Project-URL: Changelog, https://github.com/DS4SD/docling-serve/blob/main/CHANGELOG.md
|
|
12
|
+
Classifier: License :: OSI Approved :: MIT License
|
|
13
|
+
Classifier: Operating System :: OS Independent
|
|
14
|
+
Classifier: Intended Audience :: Developers
|
|
15
|
+
Classifier: Typing :: Typed
|
|
16
|
+
Classifier: Programming Language :: Python :: 3
|
|
17
|
+
Requires-Python: >=3.10
|
|
18
|
+
Description-Content-Type: text/markdown
|
|
19
|
+
License-File: LICENSE
|
|
20
|
+
Requires-Dist: docling~=2.23
|
|
21
|
+
Requires-Dist: fastapi[standard]~=0.115
|
|
22
|
+
Requires-Dist: httpx~=0.28
|
|
23
|
+
Requires-Dist: pydantic~=2.10
|
|
24
|
+
Requires-Dist: pydantic-settings~=2.4
|
|
25
|
+
Requires-Dist: python-multipart<0.1.0,>=0.0.14
|
|
26
|
+
Requires-Dist: typer~=0.12
|
|
27
|
+
Requires-Dist: uvicorn[standard]<1.0.0,>=0.29.0
|
|
28
|
+
Provides-Extra: ui
|
|
29
|
+
Requires-Dist: gradio~=5.9; extra == "ui"
|
|
30
|
+
Provides-Extra: tesserocr
|
|
31
|
+
Requires-Dist: tesserocr~=2.7; extra == "tesserocr"
|
|
32
|
+
Provides-Extra: rapidocr
|
|
33
|
+
Requires-Dist: rapidocr-onnxruntime~=1.4; python_version < "3.13" and extra == "rapidocr"
|
|
34
|
+
Requires-Dist: onnxruntime~=1.7; extra == "rapidocr"
|
|
35
|
+
Provides-Extra: cpu
|
|
36
|
+
Requires-Dist: torch>=2.6.0; extra == "cpu"
|
|
37
|
+
Requires-Dist: torchvision>=0.21.0; extra == "cpu"
|
|
38
|
+
Provides-Extra: cu124
|
|
39
|
+
Requires-Dist: torch>=2.6.0; extra == "cu124"
|
|
40
|
+
Requires-Dist: torchvision>=0.21.0; extra == "cu124"
|
|
41
|
+
|
|
42
|
+
# Docling Serve
|
|
43
|
+
|
|
44
|
+
Running [Docling](https://github.com/DS4SD/docling) as an API service.
|
|
45
|
+
|
|
46
|
+
## Usage
|
|
47
|
+
|
|
48
|
+
The API provides two endpoints: one for urls, one for files. This is necessary to send files directly in binary format instead of base64-encoded strings.
|
|
49
|
+
|
|
50
|
+
### Common parameters
|
|
51
|
+
|
|
52
|
+
On top of the source of file (see below), both endpoints support the same parameters, which are almost the same as the Docling CLI.
|
|
53
|
+
|
|
54
|
+
- `from_format` (List[str]): Input format(s) to convert from. Allowed values: `docx`, `pptx`, `html`, `image`, `pdf`, `asciidoc`, `md`. Defaults to all formats.
|
|
55
|
+
- `to_formats` (List[str]): Output format(s) to convert to. Allowed values: `md`, `json`, `html`, `text`, `doctags`. Defaults to `md`.
|
|
56
|
+
- `do_ocr` (bool): If enabled, the bitmap content will be processed using OCR. Defaults to `True`.
|
|
57
|
+
- `image_export_mode`: Image export mode for the document (only in case of JSON, Markdown or HTML). Allowed values: embedded, placeholder, referenced. Optional, defaults to `embedded`.
|
|
58
|
+
- `force_ocr` (bool): If enabled, replace any existing text with OCR-generated text over the full content. Defaults to `False`.
|
|
59
|
+
- `ocr_engine` (str): OCR engine to use. Allowed values: `easyocr`, `tesseract_cli`, `tesseract`, `rapidocr`, `ocrmac`. Defaults to `easyocr`.
|
|
60
|
+
- `ocr_lang` (List[str]): List of languages used by the OCR engine. Note that each OCR engine has different values for the language names. Defaults to empty.
|
|
61
|
+
- `pdf_backend` (str): PDF backend to use. Allowed values: `pypdfium2`, `dlparse_v1`, `dlparse_v2`. Defaults to `dlparse_v2`.
|
|
62
|
+
- `table_mode` (str): Table mode to use. Allowed values: `fast`, `accurate`. Defaults to `fast`.
|
|
63
|
+
- `abort_on_error` (bool): If enabled, abort on error. Defaults to false.
|
|
64
|
+
- `return_as_file` (boo): If enabled, return the output as a file. Defaults to false.
|
|
65
|
+
- `do_table_structure` (bool): If enabled, the table structure will be extracted. Defaults to true.
|
|
66
|
+
- `include_images` (bool): If enabled, images will be extracted from the document. Defaults to true.
|
|
67
|
+
- `images_scale` (float): Scale factor for images. Defaults to 2.0.
|
|
68
|
+
|
|
69
|
+
### URL endpoint
|
|
70
|
+
|
|
71
|
+
The endpoint is `/v1alpha/convert/source`, listening for POST requests of JSON payloads.
|
|
72
|
+
|
|
73
|
+
On top of the above parameters, you must send the URL(s) of the document you want process with either the `http_sources` or `file_sources` fields.
|
|
74
|
+
The first is fetching URL(s) (optionally using with extra headers), the second allows to provide documents as base64-encoded strings.
|
|
75
|
+
No `options` is required, they can be partially or completely omitted.
|
|
76
|
+
|
|
77
|
+
Simple payload example:
|
|
78
|
+
|
|
79
|
+
```json
|
|
80
|
+
{
|
|
81
|
+
"http_sources": [{"url": "https://arxiv.org/pdf/2206.01062"}]
|
|
82
|
+
}
|
|
83
|
+
```
|
|
84
|
+
|
|
85
|
+
<details>
|
|
86
|
+
|
|
87
|
+
<summary>Complete payload example:</summary>
|
|
88
|
+
|
|
89
|
+
```json
|
|
90
|
+
{
|
|
91
|
+
"options": {
|
|
92
|
+
"from_formats": ["docx", "pptx", "html", "image", "pdf", "asciidoc", "md", "xlsx"],
|
|
93
|
+
"to_formats": ["md", "json", "html", "text", "doctags"],
|
|
94
|
+
"image_export_mode": "placeholder",
|
|
95
|
+
"do_ocr": true,
|
|
96
|
+
"force_ocr": false,
|
|
97
|
+
"ocr_engine": "easyocr",
|
|
98
|
+
"ocr_lang": ["en"],
|
|
99
|
+
"pdf_backend": "dlparse_v2",
|
|
100
|
+
"table_mode": "fast",
|
|
101
|
+
"abort_on_error": false,
|
|
102
|
+
"return_as_file": false,
|
|
103
|
+
},
|
|
104
|
+
"http_sources": [{"url": "https://arxiv.org/pdf/2206.01062"}]
|
|
105
|
+
}
|
|
106
|
+
```
|
|
107
|
+
|
|
108
|
+
</details>
|
|
109
|
+
|
|
110
|
+
<details>
|
|
111
|
+
|
|
112
|
+
<summary>CURL example:</summary>
|
|
113
|
+
|
|
114
|
+
```sh
|
|
115
|
+
curl -X 'POST' \
|
|
116
|
+
'http://localhost:5001/v1alpha/convert/source' \
|
|
117
|
+
-H 'accept: application/json' \
|
|
118
|
+
-H 'Content-Type: application/json' \
|
|
119
|
+
-d '{
|
|
120
|
+
"options": {
|
|
121
|
+
"from_formats": [
|
|
122
|
+
"docx",
|
|
123
|
+
"pptx",
|
|
124
|
+
"html",
|
|
125
|
+
"image",
|
|
126
|
+
"pdf",
|
|
127
|
+
"asciidoc",
|
|
128
|
+
"md",
|
|
129
|
+
"xlsx"
|
|
130
|
+
],
|
|
131
|
+
"to_formats": ["md", "json", "html", "text", "doctags"],
|
|
132
|
+
"image_export_mode": "placeholder",
|
|
133
|
+
"do_ocr": true,
|
|
134
|
+
"force_ocr": false,
|
|
135
|
+
"ocr_engine": "easyocr",
|
|
136
|
+
"ocr_lang": [
|
|
137
|
+
"fr",
|
|
138
|
+
"de",
|
|
139
|
+
"es",
|
|
140
|
+
"en"
|
|
141
|
+
],
|
|
142
|
+
"pdf_backend": "dlparse_v2",
|
|
143
|
+
"table_mode": "fast",
|
|
144
|
+
"abort_on_error": false,
|
|
145
|
+
"return_as_file": false,
|
|
146
|
+
"do_table_structure": true,
|
|
147
|
+
"include_images": true,
|
|
148
|
+
"images_scale": 2,
|
|
149
|
+
},
|
|
150
|
+
"http_sources": [{"url": "https://arxiv.org/pdf/2206.01062"}]
|
|
151
|
+
}'
|
|
152
|
+
```
|
|
153
|
+
|
|
154
|
+
</details>
|
|
155
|
+
|
|
156
|
+
<details>
|
|
157
|
+
<summary>Python example:</summary>
|
|
158
|
+
|
|
159
|
+
```python
|
|
160
|
+
import httpx
|
|
161
|
+
|
|
162
|
+
async_client = httpx.AsyncClient(timeout=60.0)
|
|
163
|
+
url = "http://localhost:5001/v1alpha/convert/source"
|
|
164
|
+
payload = {
|
|
165
|
+
"options": {
|
|
166
|
+
"from_formats": ["docx", "pptx", "html", "image", "pdf", "asciidoc", "md", "xlsx"],
|
|
167
|
+
"to_formats": ["md", "json", "html", "text", "doctags"],
|
|
168
|
+
"image_export_mode": "placeholder",
|
|
169
|
+
"do_ocr": True,
|
|
170
|
+
"force_ocr": False,
|
|
171
|
+
"ocr_engine": "easyocr",
|
|
172
|
+
"ocr_lang": "en",
|
|
173
|
+
"pdf_backend": "dlparse_v2",
|
|
174
|
+
"table_mode": "fast",
|
|
175
|
+
"abort_on_error": False,
|
|
176
|
+
"return_as_file": False,
|
|
177
|
+
},
|
|
178
|
+
"http_sources": [{"url": "https://arxiv.org/pdf/2206.01062"}]
|
|
179
|
+
}
|
|
180
|
+
|
|
181
|
+
response = await async_client_client.post(url, json=payload)
|
|
182
|
+
|
|
183
|
+
data = response.json()
|
|
184
|
+
```
|
|
185
|
+
|
|
186
|
+
</details>
|
|
187
|
+
|
|
188
|
+
#### File as base64
|
|
189
|
+
|
|
190
|
+
The `file_sources` argument in the endpoint allows to send files as base64-encoded strings.
|
|
191
|
+
When your PDF or other file type is too large, encoding it and passing it inline to curl
|
|
192
|
+
can lead to an โArgument list too longโ error on some systems. To avoid this, we write
|
|
193
|
+
the JSON request body to a file and have curl read from that file.
|
|
194
|
+
|
|
195
|
+
<details>
|
|
196
|
+
<summary>CURL steps:</summary>
|
|
197
|
+
|
|
198
|
+
```sh
|
|
199
|
+
# 1. Base64-encode the file
|
|
200
|
+
B64_DATA=$(base64 -w 0 /path/to/file/pdf-to-convert.pdf)
|
|
201
|
+
|
|
202
|
+
# 2. Build the JSON with your options
|
|
203
|
+
cat <<EOF > /tmp/request_body.json
|
|
204
|
+
{
|
|
205
|
+
"options": {
|
|
206
|
+
},
|
|
207
|
+
"file_sources": [{
|
|
208
|
+
"base64_string": "${B64_DATA}",
|
|
209
|
+
"filename": "pdf-to-convert.pdf"
|
|
210
|
+
}]
|
|
211
|
+
}
|
|
212
|
+
EOF
|
|
213
|
+
|
|
214
|
+
# 3. POST the request to the docling service
|
|
215
|
+
curl -X POST "localhost:5001/v1alpha/convert/source" \
|
|
216
|
+
-H "Content-Type: application/json" \
|
|
217
|
+
-d @/tmp/request_body.json
|
|
218
|
+
```
|
|
219
|
+
|
|
220
|
+
</details>
|
|
221
|
+
|
|
222
|
+
### File endpoint
|
|
223
|
+
|
|
224
|
+
The endpoint is: `/v1alpha/convert/file`, listening for POST requests of Form payloads (necessary as the files are sent as multipart/form data). You can send one or multiple files.
|
|
225
|
+
|
|
226
|
+
<details>
|
|
227
|
+
<summary>CURL example:</summary>
|
|
228
|
+
|
|
229
|
+
```sh
|
|
230
|
+
curl -X 'POST' \
|
|
231
|
+
'http://127.0.0.1:5001/v1alpha/convert/file' \
|
|
232
|
+
-H 'accept: application/json' \
|
|
233
|
+
-H 'Content-Type: multipart/form-data' \
|
|
234
|
+
-F 'ocr_engine=easyocr' \
|
|
235
|
+
-F 'pdf_backend=dlparse_v2' \
|
|
236
|
+
-F 'from_formats=pdf' \
|
|
237
|
+
-F 'from_formats=docx' \
|
|
238
|
+
-F 'force_ocr=false' \
|
|
239
|
+
-F 'image_export_mode=embedded' \
|
|
240
|
+
-F 'ocr_lang=en' \
|
|
241
|
+
-F 'ocr_lang=pl' \
|
|
242
|
+
-F 'table_mode=fast' \
|
|
243
|
+
-F 'files=@2206.01062v1.pdf;type=application/pdf' \
|
|
244
|
+
-F 'abort_on_error=false' \
|
|
245
|
+
-F 'to_formats=md' \
|
|
246
|
+
-F 'to_formats=text' \
|
|
247
|
+
-F 'return_as_file=false' \
|
|
248
|
+
-F 'do_ocr=true'
|
|
249
|
+
```
|
|
250
|
+
|
|
251
|
+
</details>
|
|
252
|
+
|
|
253
|
+
<details>
|
|
254
|
+
<summary>Python example:</summary>
|
|
255
|
+
|
|
256
|
+
```python
|
|
257
|
+
import httpx
|
|
258
|
+
|
|
259
|
+
async_client = httpx.AsyncClient(timeout=60.0)
|
|
260
|
+
url = "http://localhost:5001/v1alpha/convert/file"
|
|
261
|
+
parameters = {
|
|
262
|
+
"from_formats": ["docx", "pptx", "html", "image", "pdf", "asciidoc", "md", "xlsx"],
|
|
263
|
+
"to_formats": ["md", "json", "html", "text", "doctags"],
|
|
264
|
+
"image_export_mode": "placeholder",
|
|
265
|
+
"do_ocr": True,
|
|
266
|
+
"force_ocr": False,
|
|
267
|
+
"ocr_engine": "easyocr",
|
|
268
|
+
"ocr_lang": ["en"],
|
|
269
|
+
"pdf_backend": "dlparse_v2",
|
|
270
|
+
"table_mode": "fast",
|
|
271
|
+
"abort_on_error": False,
|
|
272
|
+
"return_as_file": False
|
|
273
|
+
}
|
|
274
|
+
|
|
275
|
+
current_dir = os.path.dirname(__file__)
|
|
276
|
+
file_path = os.path.join(current_dir, '2206.01062v1.pdf')
|
|
277
|
+
|
|
278
|
+
files = {
|
|
279
|
+
'files': ('2206.01062v1.pdf', open(file_path, 'rb'), 'application/pdf'),
|
|
280
|
+
}
|
|
281
|
+
|
|
282
|
+
response = await async_client.post(url, files=files, data={"parameters": json.dumps(parameters)})
|
|
283
|
+
assert response.status_code == 200, "Response should be 200 OK"
|
|
284
|
+
|
|
285
|
+
data = response.json()
|
|
286
|
+
```
|
|
287
|
+
|
|
288
|
+
</details>
|
|
289
|
+
|
|
290
|
+
### Response format
|
|
291
|
+
|
|
292
|
+
The response can be a JSON Document or a File.
|
|
293
|
+
|
|
294
|
+
- If you process only one file, the response will be a JSON document with the following format:
|
|
295
|
+
|
|
296
|
+
```jsonc
|
|
297
|
+
{
|
|
298
|
+
"document": {
|
|
299
|
+
"md_content": "",
|
|
300
|
+
"json_content": {},
|
|
301
|
+
"html_content": "",
|
|
302
|
+
"text_content": "",
|
|
303
|
+
"doctags_content": ""
|
|
304
|
+
},
|
|
305
|
+
"status": "<success|partial_success|skipped|failure>",
|
|
306
|
+
"processing_time": 0.0,
|
|
307
|
+
"timings": {},
|
|
308
|
+
"errors": []
|
|
309
|
+
}
|
|
310
|
+
```
|
|
311
|
+
|
|
312
|
+
Depending on the value you set in `output_formats`, the different items will be populated with their respective results or empty.
|
|
313
|
+
|
|
314
|
+
`processing_time` is the Docling processing time in seconds, and `timings` (when enabled in the backend) provides the detailed
|
|
315
|
+
timing of all the internal Docling components.
|
|
316
|
+
|
|
317
|
+
- If you set the parameter `return_as_file` to True, the response will be a zip file.
|
|
318
|
+
- If multiple files are generated (multiple inputs, or one input but multiple outputs with `return_as_file` True), the response will be a zip file.
|
|
319
|
+
|
|
320
|
+
## Helpers
|
|
321
|
+
|
|
322
|
+
- A full Swagger UI is available at the `/docs` endpoint.
|
|
323
|
+
|
|
324
|
+

|
|
325
|
+
|
|
326
|
+
- An easy to use UI is available at the `/ui` endpoint.
|
|
327
|
+
|
|
328
|
+

|
|
329
|
+
|
|
330
|
+

|
|
331
|
+
|
|
332
|
+
## Development
|
|
333
|
+
|
|
334
|
+
### CPU only
|
|
335
|
+
|
|
336
|
+
```sh
|
|
337
|
+
# Install uv if not already available
|
|
338
|
+
curl -LsSf https://astral.sh/uv/install.sh | sh
|
|
339
|
+
|
|
340
|
+
# Install dependencies
|
|
341
|
+
uv sync --extra cpu
|
|
342
|
+
```
|
|
343
|
+
|
|
344
|
+
### Cuda GPU
|
|
345
|
+
|
|
346
|
+
For GPU support use the following command:
|
|
347
|
+
|
|
348
|
+
```sh
|
|
349
|
+
# Install dependencies
|
|
350
|
+
uv sync
|
|
351
|
+
```
|
|
352
|
+
|
|
353
|
+
### Gradio UI and different OCR backends
|
|
354
|
+
|
|
355
|
+
`/ui` endpoint using `gradio` and different OCR backends can be enabled via package extras:
|
|
356
|
+
|
|
357
|
+
```sh
|
|
358
|
+
# Enable ui and rapidocr
|
|
359
|
+
uv sync --extra ui --extra rapidocr
|
|
360
|
+
```
|
|
361
|
+
|
|
362
|
+
```sh
|
|
363
|
+
# Enable tesserocr
|
|
364
|
+
uv sync --extra tesserocr
|
|
365
|
+
```
|
|
366
|
+
|
|
367
|
+
See `[project.optional-dependencies]` section in `pyproject.toml` for full list of options.
|
|
368
|
+
|
|
369
|
+
### Run the server
|
|
370
|
+
|
|
371
|
+
The `docling-serve` executable is a convenient script for launching the webserver both in
|
|
372
|
+
development and production mode.
|
|
373
|
+
|
|
374
|
+
```sh
|
|
375
|
+
# Run the server in development mode
|
|
376
|
+
# - reload is enabled by default
|
|
377
|
+
# - listening on the 127.0.0.1 address
|
|
378
|
+
# - ui is enabled by default
|
|
379
|
+
docling-serve dev
|
|
380
|
+
|
|
381
|
+
# Run the server in production mode
|
|
382
|
+
# - reload is disabled by default
|
|
383
|
+
# - listening on the 0.0.0.0 address
|
|
384
|
+
# - ui is disabled by default
|
|
385
|
+
docling-serve run
|
|
386
|
+
```
|
|
387
|
+
|
|
388
|
+
### Options
|
|
389
|
+
|
|
390
|
+
The `docling-serve` executable allows is controlled with both command line
|
|
391
|
+
options and environment variables.
|
|
392
|
+
|
|
393
|
+
<details>
|
|
394
|
+
<summary>`docling-serve` help message</summary>
|
|
395
|
+
|
|
396
|
+
```sh
|
|
397
|
+
$ docling-serve dev --help
|
|
398
|
+
|
|
399
|
+
Usage: docling-serve dev [OPTIONS]
|
|
400
|
+
|
|
401
|
+
Run a Docling Serve app in development mode. ๐งช
|
|
402
|
+
This is equivalent to docling-serve run but with reload
|
|
403
|
+
enabled and listening on the 127.0.0.1 address.
|
|
404
|
+
|
|
405
|
+
Options can be set also with the corresponding ENV variable, with the exception
|
|
406
|
+
of --enable-ui, --host and --reload.
|
|
407
|
+
|
|
408
|
+
โญโ Options โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฎ
|
|
409
|
+
โ --host TEXT The host to serve on. For local development in localhost โ
|
|
410
|
+
โ use 127.0.0.1. To enable public access, e.g. in a โ
|
|
411
|
+
โ container, use all the IP addresses available with โ
|
|
412
|
+
โ 0.0.0.0. โ
|
|
413
|
+
โ [default: 127.0.0.1] โ
|
|
414
|
+
โ --port INTEGER The port to serve on. [default: 5001] โ
|
|
415
|
+
โ --reload --no-reload Enable auto-reload of the server when (code) files โ
|
|
416
|
+
โ change. This is resource intensive, use it only during โ
|
|
417
|
+
โ development. โ
|
|
418
|
+
โ [default: reload] โ
|
|
419
|
+
โ --root-path TEXT The root path is used to tell your app that it is being โ
|
|
420
|
+
โ served to the outside world with some path prefix set up โ
|
|
421
|
+
โ in some termination proxy or similar. โ
|
|
422
|
+
โ --proxy-headers --no-proxy-headers Enable/Disable X-Forwarded-Proto, X-Forwarded-For, โ
|
|
423
|
+
โ X-Forwarded-Port to populate remote address info. โ
|
|
424
|
+
โ [default: proxy-headers] โ
|
|
425
|
+
โ --artifacts-path PATH If set to a valid directory, the model weights will be โ
|
|
426
|
+
โ loaded from this path. โ
|
|
427
|
+
โ [default: None] โ
|
|
428
|
+
โ --enable-ui --no-enable-ui Enable the development UI. [default: enable-ui] โ
|
|
429
|
+
โ --help Show this message and exit. โ
|
|
430
|
+
โฐโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฏ
|
|
431
|
+
```
|
|
432
|
+
|
|
433
|
+
</details>
|
|
434
|
+
|
|
435
|
+
#### Environment variables
|
|
436
|
+
|
|
437
|
+
The environment variables controlling the `uvicorn` execution can be specified with the `UVICORN_` prefix:
|
|
438
|
+
|
|
439
|
+
- `UVICORN_WORKERS`: Number of workers to use.
|
|
440
|
+
- `UVICORN_RELOAD`: If `True`, this will enable auto-reload when you modify files, useful for development.
|
|
441
|
+
|
|
442
|
+
The environment variables controlling specifics of the Docling Serve app can be specified with the
|
|
443
|
+
`DOCLING_SERVE_` prefix:
|
|
444
|
+
|
|
445
|
+
- `DOCLING_SERVE_ARTIFACTS_PATH`: if set Docling will use only the local weights of models, for example `/opt/app-root/src/.cache/docling/models`.
|
|
446
|
+
- `DOCLING_SERVE_ENABLE_UI`: If `True`, The Gradio UI will be available at `/ui`.
|
|
447
|
+
|
|
448
|
+
Others:
|
|
449
|
+
|
|
450
|
+
- `TESSDATA_PREFIX`: Tesseract data location, example `/usr/share/tesseract/tessdata/`.
|
|
451
|
+
|
|
452
|
+
## Get help and support
|
|
453
|
+
|
|
454
|
+
Please feel free to connect with us using the [discussion section](https://github.com/DS4SD/docling/discussions).
|
|
455
|
+
|
|
456
|
+
## Contributing
|
|
457
|
+
|
|
458
|
+
Please read [Contributing to Docling Serve](https://github.com/DS4SD/docling-serve/blob/main/CONTRIBUTING.md) for details.
|
|
459
|
+
|
|
460
|
+
## References
|
|
461
|
+
|
|
462
|
+
If you use Docling in your projects, please consider citing the following:
|
|
463
|
+
|
|
464
|
+
```bib
|
|
465
|
+
@techreport{Docling,
|
|
466
|
+
author = {Deep Search Team},
|
|
467
|
+
month = {8},
|
|
468
|
+
title = {Docling Technical Report},
|
|
469
|
+
url = {https://arxiv.org/abs/2408.09869},
|
|
470
|
+
eprint = {2408.09869},
|
|
471
|
+
doi = {10.48550/arXiv.2408.09869},
|
|
472
|
+
version = {1.0.0},
|
|
473
|
+
year = {2024}
|
|
474
|
+
}
|
|
475
|
+
```
|
|
476
|
+
|
|
477
|
+
## License
|
|
478
|
+
|
|
479
|
+
The Docling Serve codebase is under MIT license.
|
|
480
|
+
|
|
481
|
+
## IBM โค๏ธ Open Source AI
|
|
482
|
+
|
|
483
|
+
Docling has been brought to you by IBM.
|