Update main.py
Browse files
main.py
CHANGED
@@ -1,51 +1,19 @@
|
|
1 |
-
try: from pip._internal.operations import freeze
|
2 |
-
except ImportError: # pip < 10.0
|
3 |
-
from pip.operations import freeze
|
4 |
-
|
5 |
-
pkgs = freeze.freeze()
|
6 |
-
for pkg in pkgs: print(pkg)
|
7 |
-
|
8 |
import io
|
9 |
import asyncio
|
10 |
-
import time
|
11 |
-
import aiohttp
|
12 |
-
from PyPDF2 import PdfReader, PdfWriter
|
13 |
import os
|
14 |
from pathlib import Path
|
15 |
-
|
16 |
-
from
|
|
|
17 |
from fastapi.responses import JSONResponse
|
18 |
-
from
|
|
|
19 |
|
20 |
app = FastAPI()
|
21 |
|
22 |
-
|
23 |
-
app.add_middleware(
|
24 |
-
CORSMiddleware,
|
25 |
-
allow_origins=["*"],
|
26 |
-
allow_credentials=True,
|
27 |
-
allow_methods=["*"],
|
28 |
-
allow_headers=["*"],
|
29 |
-
)
|
30 |
-
|
31 |
-
async def execute_pdfscraper_async(file_contents: bytes, file_name: str, pages_per_chunk: int):
|
32 |
-
split_pdf(file_contents, file_name, pages_per_chunk)
|
33 |
-
response_list = []
|
34 |
-
async with aiohttp.ClientSession() as session:
|
35 |
-
tasks = [
|
36 |
-
call_pdfscraper(session, chunk_data, chunk_name)
|
37 |
-
for chunk_data, chunk_name in load_chunks(file_name)
|
38 |
-
]
|
39 |
-
responses = await asyncio.gather(*tasks)
|
40 |
-
for response in responses:
|
41 |
-
response_list.append(response[0])
|
42 |
-
|
43 |
-
return response_list
|
44 |
-
|
45 |
-
async def call_pdfscraper(session, file_contents, pdf_name):
|
46 |
headers = {"Origin": "http://localhost:8080"}
|
47 |
url = "https://us-central1-neuralgap-1.cloudfunctions.net/scraperPDFDocxTables_v3"
|
48 |
-
# Create a FormData object
|
49 |
data = FormData()
|
50 |
data.add_field(
|
51 |
"pdf",
|
@@ -53,24 +21,42 @@ async def call_pdfscraper(session, file_contents, pdf_name):
|
|
53 |
filename=os.path.basename(pdf_name),
|
54 |
content_type="application/pdf",
|
55 |
)
|
56 |
-
data.add_field("processTables",
|
57 |
|
58 |
async with session.post(url, data=data, headers=headers) as resp:
|
59 |
if resp.status == 200:
|
60 |
response = await resp.json()
|
61 |
else:
|
62 |
-
|
63 |
-
return {}
|
64 |
|
65 |
return response, pdf_name
|
66 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
67 |
def collect_pdfscraper_response(scrape_response_list):
|
68 |
content_list = []
|
69 |
tables_dict = {}
|
70 |
table_count = 1
|
71 |
for response in scrape_response_list:
|
72 |
-
content = response
|
73 |
-
table_content = response
|
74 |
|
75 |
content_list.append(content)
|
76 |
try:
|
@@ -85,11 +71,11 @@ def collect_pdfscraper_response(scrape_response_list):
|
|
85 |
|
86 |
return content_str, tables_dict
|
87 |
|
|
|
88 |
def split_pdf(file_contents, file_name, pages_per_chunk):
|
89 |
file_bytes = io.BytesIO(file_contents)
|
90 |
reader = PdfReader(file_bytes)
|
91 |
total_pages = len(reader.pages)
|
92 |
-
|
93 |
output_dir = Path(file_name).parent / "chunks"
|
94 |
os.makedirs(output_dir, exist_ok=True)
|
95 |
|
@@ -108,27 +94,23 @@ def split_pdf(file_contents, file_name, pages_per_chunk):
|
|
108 |
with open(output_path, "wb") as output_pdf:
|
109 |
writer.write(output_pdf)
|
110 |
|
111 |
-
|
112 |
-
|
113 |
-
chunk_list = os.listdir(output_dir)
|
114 |
-
chunk_byte_list = [
|
115 |
-
(open(f"{output_dir}/{file}", "rb").read(), file) for file in chunk_list
|
116 |
-
]
|
117 |
-
return chunk_byte_list
|
118 |
|
119 |
@app.post("/process-pdf/")
|
120 |
-
async def process_pdf(
|
121 |
-
file_contents = await
|
122 |
-
file_name = file.filename
|
123 |
|
124 |
-
|
|
|
125 |
content, table_string = collect_pdfscraper_response(scrape_response_list)
|
126 |
-
|
127 |
-
shutil.rmtree(
|
128 |
-
|
129 |
return JSONResponse(content={"content": content, "tables": table_string})
|
130 |
|
131 |
-
|
|
|
132 |
# if __name__ == "__main__":
|
133 |
# import uvicorn
|
134 |
-
# uvicorn.run(app,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
import io
|
2 |
import asyncio
|
|
|
|
|
|
|
3 |
import os
|
4 |
from pathlib import Path
|
5 |
+
import aiohttp
|
6 |
+
from PyPDF2 import PdfReader, PdfWriter
|
7 |
+
from fastapi import FastAPI, UploadFile, Form
|
8 |
from fastapi.responses import JSONResponse
|
9 |
+
from aiohttp import FormData
|
10 |
+
import shutil
|
11 |
|
12 |
app = FastAPI()
|
13 |
|
14 |
+
async def call_pdfscraper(session, file_contents, pdf_name, processTables):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
15 |
headers = {"Origin": "http://localhost:8080"}
|
16 |
url = "https://us-central1-neuralgap-1.cloudfunctions.net/scraperPDFDocxTables_v3"
|
|
|
17 |
data = FormData()
|
18 |
data.add_field(
|
19 |
"pdf",
|
|
|
21 |
filename=os.path.basename(pdf_name),
|
22 |
content_type="application/pdf",
|
23 |
)
|
24 |
+
data.add_field("processTables", processTables)
|
25 |
|
26 |
async with session.post(url, data=data, headers=headers) as resp:
|
27 |
if resp.status == 200:
|
28 |
response = await resp.json()
|
29 |
else:
|
30 |
+
return {}, pdf_name
|
|
|
31 |
|
32 |
return response, pdf_name
|
33 |
|
34 |
+
|
35 |
+
async def execute_pdfscraper_async(file_path: str, processTables: str):
|
36 |
+
chunk_list = os.listdir(file_path)
|
37 |
+
chunk_byte_list = [
|
38 |
+
(open(f"{file_path}/{file}", "rb").read(), file) for file in chunk_list
|
39 |
+
]
|
40 |
+
response_list = []
|
41 |
+
async with aiohttp.ClientSession() as session:
|
42 |
+
tasks = [
|
43 |
+
call_pdfscraper(session, file_all[0], file_all[1], processTables)
|
44 |
+
for file_all in chunk_byte_list
|
45 |
+
]
|
46 |
+
responses = await asyncio.gather(*tasks)
|
47 |
+
for i, response in enumerate(responses):
|
48 |
+
response_list.append(response[0])
|
49 |
+
|
50 |
+
return response_list
|
51 |
+
|
52 |
+
|
53 |
def collect_pdfscraper_response(scrape_response_list):
|
54 |
content_list = []
|
55 |
tables_dict = {}
|
56 |
table_count = 1
|
57 |
for response in scrape_response_list:
|
58 |
+
content = response.get("corpus", "")
|
59 |
+
table_content = response.get("tables_raw", {})
|
60 |
|
61 |
content_list.append(content)
|
62 |
try:
|
|
|
71 |
|
72 |
return content_str, tables_dict
|
73 |
|
74 |
+
|
75 |
def split_pdf(file_contents, file_name, pages_per_chunk):
|
76 |
file_bytes = io.BytesIO(file_contents)
|
77 |
reader = PdfReader(file_bytes)
|
78 |
total_pages = len(reader.pages)
|
|
|
79 |
output_dir = Path(file_name).parent / "chunks"
|
80 |
os.makedirs(output_dir, exist_ok=True)
|
81 |
|
|
|
94 |
with open(output_path, "wb") as output_pdf:
|
95 |
writer.write(output_pdf)
|
96 |
|
97 |
+
return str(output_dir)
|
98 |
+
|
|
|
|
|
|
|
|
|
|
|
99 |
|
100 |
@app.post("/process-pdf/")
|
101 |
+
async def process_pdf(pdf_file: UploadFile, pages_per_chunk: int = Form(2), processTables: str = Form("True")):
|
102 |
+
file_contents = await pdf_file.read()
|
|
|
103 |
|
104 |
+
chunks_dir = split_pdf(file_contents, pdf_file.filename, pages_per_chunk)
|
105 |
+
scrape_response_list = await execute_pdfscraper_async(chunks_dir, processTables)
|
106 |
content, table_string = collect_pdfscraper_response(scrape_response_list)
|
107 |
+
|
108 |
+
shutil.rmtree(chunks_dir) # Clean up chunks after processing
|
109 |
+
|
110 |
return JSONResponse(content={"content": content, "tables": table_string})
|
111 |
|
112 |
+
|
113 |
+
# Starting point for running the FastAPI app
|
114 |
# if __name__ == "__main__":
|
115 |
# import uvicorn
|
116 |
+
# uvicorn.run(app, host="0.0.0.0", port=8000)
|