update the queue
Browse files
__pycache__/inference_svm_model.cpython-310.pyc
DELETED
Binary file (960 Bytes)
|
|
__pycache__/mineru_single.cpython-310.pyc
DELETED
Binary file (4.24 kB)
|
|
__pycache__/worker.cpython-310.pyc
DELETED
Binary file (4.85 kB)
|
|
app.py
CHANGED
@@ -42,7 +42,7 @@ async def process_pdf(
|
|
42 |
|
43 |
channel.basic_publish(
|
44 |
exchange="",
|
45 |
-
routing_key="
|
46 |
body=json.dumps(input_json),
|
47 |
properties=pika.BasicProperties(
|
48 |
headers={"process": "topic_extraction"}
|
|
|
42 |
|
43 |
channel.basic_publish(
|
44 |
exchange="",
|
45 |
+
routing_key="gpu_server",
|
46 |
body=json.dumps(input_json),
|
47 |
properties=pika.BasicProperties(
|
48 |
headers={"process": "topic_extraction"}
|
download_models_hf.py
CHANGED
@@ -1,67 +1,66 @@
|
|
1 |
-
import json
|
2 |
-
import os
|
3 |
-
|
4 |
-
import requests
|
5 |
-
from huggingface_hub import snapshot_download
|
6 |
-
|
7 |
-
|
8 |
-
def download_json(url):
|
9 |
-
# 下载JSON文件
|
10 |
-
response = requests.get(url)
|
11 |
-
response.raise_for_status() # 检查请求是否成功
|
12 |
-
return response.json()
|
13 |
-
|
14 |
-
|
15 |
-
def download_and_modify_json(url, local_filename, modifications):
|
16 |
-
if os.path.exists(local_filename):
|
17 |
-
data = json.load(open(local_filename))
|
18 |
-
config_version = data.get('config_version', '0.0.0')
|
19 |
-
if config_version < '1.1.1':
|
20 |
-
data = download_json(url)
|
21 |
-
else:
|
22 |
-
data = download_json(url)
|
23 |
-
|
24 |
-
# 修改内容
|
25 |
-
for key, value in modifications.items():
|
26 |
-
data[key] = value
|
27 |
-
|
28 |
-
# 保存修改后的内容
|
29 |
-
with open(local_filename, 'w', encoding='utf-8') as f:
|
30 |
-
json.dump(data, f, ensure_ascii=False, indent=4)
|
31 |
-
|
32 |
-
|
33 |
-
if __name__ == '__main__':
|
34 |
-
|
35 |
-
mineru_patterns = [
|
36 |
-
"models/Layout/LayoutLMv3/*",
|
37 |
-
"models/Layout/YOLO/*",
|
38 |
-
"models/MFD/YOLO/*",
|
39 |
-
"models/MFR/unimernet_small_2501/*",
|
40 |
-
"models/TabRec/TableMaster/*",
|
41 |
-
"models/TabRec/StructEqTable/*",
|
42 |
-
]
|
43 |
-
model_dir = snapshot_download('opendatalab/PDF-Extract-Kit-1.0', allow_patterns=mineru_patterns)
|
44 |
-
|
45 |
-
layoutreader_pattern = [
|
46 |
-
"*.json",
|
47 |
-
"*.safetensors",
|
48 |
-
]
|
49 |
-
layoutreader_model_dir = snapshot_download('hantian/layoutreader', allow_patterns=layoutreader_pattern)
|
50 |
-
|
51 |
-
model_dir = model_dir + '/models'
|
52 |
-
print(f'model_dir is: {model_dir}')
|
53 |
-
print(f'layoutreader_model_dir is: {layoutreader_model_dir}')
|
54 |
-
|
55 |
-
json_url = 'https://github.com/opendatalab/MinerU/raw/master/magic-pdf.template.json'
|
56 |
-
config_file_name = 'magic-pdf.json'
|
57 |
-
home_dir = os.path.expanduser('~')
|
58 |
-
config_file = os.path.join(home_dir, config_file_name)
|
59 |
-
|
60 |
-
json_mods = {
|
61 |
-
'models-dir': model_dir,
|
62 |
-
'layoutreader-model-dir': layoutreader_model_dir,
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
print(f'The configuration file has been configured successfully, the path is: {config_file}')
|
|
|
1 |
+
import json
|
2 |
+
import os
|
3 |
+
|
4 |
+
import requests
|
5 |
+
from huggingface_hub import snapshot_download
|
6 |
+
|
7 |
+
|
8 |
+
def download_json(url):
|
9 |
+
# 下载JSON文件
|
10 |
+
response = requests.get(url)
|
11 |
+
response.raise_for_status() # 检查请求是否成功
|
12 |
+
return response.json()
|
13 |
+
|
14 |
+
|
15 |
+
def download_and_modify_json(url, local_filename, modifications):
|
16 |
+
if os.path.exists(local_filename):
|
17 |
+
data = json.load(open(local_filename))
|
18 |
+
config_version = data.get('config_version', '0.0.0')
|
19 |
+
if config_version < '1.1.1':
|
20 |
+
data = download_json(url)
|
21 |
+
else:
|
22 |
+
data = download_json(url)
|
23 |
+
|
24 |
+
# 修改内容
|
25 |
+
for key, value in modifications.items():
|
26 |
+
data[key] = value
|
27 |
+
|
28 |
+
# 保存修改后的内容
|
29 |
+
with open(local_filename, 'w', encoding='utf-8') as f:
|
30 |
+
json.dump(data, f, ensure_ascii=False, indent=4)
|
31 |
+
|
32 |
+
|
33 |
+
if __name__ == '__main__':
|
34 |
+
|
35 |
+
mineru_patterns = [
|
36 |
+
"models/Layout/LayoutLMv3/*",
|
37 |
+
"models/Layout/YOLO/*",
|
38 |
+
"models/MFD/YOLO/*",
|
39 |
+
"models/MFR/unimernet_small_2501/*",
|
40 |
+
"models/TabRec/TableMaster/*",
|
41 |
+
"models/TabRec/StructEqTable/*",
|
42 |
+
]
|
43 |
+
model_dir = snapshot_download('opendatalab/PDF-Extract-Kit-1.0', allow_patterns=mineru_patterns)
|
44 |
+
|
45 |
+
layoutreader_pattern = [
|
46 |
+
"*.json",
|
47 |
+
"*.safetensors",
|
48 |
+
]
|
49 |
+
layoutreader_model_dir = snapshot_download('hantian/layoutreader', allow_patterns=layoutreader_pattern)
|
50 |
+
|
51 |
+
model_dir = model_dir + '/models'
|
52 |
+
print(f'model_dir is: {model_dir}')
|
53 |
+
print(f'layoutreader_model_dir is: {layoutreader_model_dir}')
|
54 |
+
|
55 |
+
json_url = 'https://github.com/opendatalab/MinerU/raw/master/magic-pdf.template.json'
|
56 |
+
config_file_name = 'magic-pdf.json'
|
57 |
+
home_dir = os.path.expanduser('~')
|
58 |
+
config_file = os.path.join(home_dir, config_file_name)
|
59 |
+
|
60 |
+
json_mods = {
|
61 |
+
'models-dir': model_dir,
|
62 |
+
'layoutreader-model-dir': layoutreader_model_dir,
|
63 |
+
}
|
64 |
+
|
65 |
+
download_and_modify_json(json_url, config_file, json_mods)
|
66 |
+
print(f'The configuration file has been configured successfully, the path is: {config_file}')
|
|