Spaces:
Sleeping
Sleeping
rudradeep22
commited on
Commit
•
9a0d2e1
1
Parent(s):
dde7c4f
panel app for hackathon
Browse files- Dockerfile +11 -0
- app.py +110 -0
- requirements.txt +4 -0
Dockerfile
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
FROM python:3.9
|
2 |
+
|
3 |
+
WORKDIR /code
|
4 |
+
|
5 |
+
COPY ./requirements.txt /code/requirements.txt
|
6 |
+
|
7 |
+
RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
|
8 |
+
|
9 |
+
COPY . .
|
10 |
+
|
11 |
+
CMD ["panel", "serve", "/code/app.py", "--address", "0.0.0.0", "--port", "7860", "--allow-websocket-origin", "InvictusRudra-question-answering.hf.space"]
|
app.py
ADDED
@@ -0,0 +1,110 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# -*- coding: utf-8 -*-
|
2 |
+
"""Hackathon_Illuminati.ipynb
|
3 |
+
|
4 |
+
Automatically generated by Colaboratory.
|
5 |
+
|
6 |
+
Original file is located at
|
7 |
+
https://colab.research.google.com/drive/1B-SaMQ85UdV9DnqZ6_OWqg8oOAlgcZ5V
|
8 |
+
|
9 |
+
# LangChain QA Panel App
|
10 |
+
|
11 |
+
This notebook shows how to make this app:
|
12 |
+
"""
|
13 |
+
|
14 |
+
# !pip install transformers
|
15 |
+
# !pip install easyocr
|
16 |
+
# !pip install pdf2image
|
17 |
+
# !apt-get install poppler-utils
|
18 |
+
#
|
19 |
+
# ! pip install PyPDF2
|
20 |
+
|
21 |
+
import panel as pn
|
22 |
+
from transformers import pipeline
|
23 |
+
from pdf2image import convert_from_path
|
24 |
+
import easyocr
|
25 |
+
|
26 |
+
pn.extension('texteditor', template="bootstrap", sizing_mode='stretch_width')
|
27 |
+
pn.state.template.param.update(
|
28 |
+
main_max_width="690px",
|
29 |
+
header_background="#F08080",
|
30 |
+
)
|
31 |
+
|
32 |
+
file_input = pn.widgets.FileInput(width=300)
|
33 |
+
|
34 |
+
prompt = pn.widgets.TextEditor(
|
35 |
+
value="", placeholder="Enter your questions here...", height=160, toolbar=False
|
36 |
+
)
|
37 |
+
run_button = pn.widgets.Button(name="Run")
|
38 |
+
|
39 |
+
widgets = pn.Row(
|
40 |
+
pn.Column(prompt, run_button, margin=5), width = 630
|
41 |
+
)
|
42 |
+
|
43 |
+
def qa(file, query):
|
44 |
+
|
45 |
+
images = convert_from_path(file)
|
46 |
+
reader = easyocr.Reader(['en'])
|
47 |
+
result = []
|
48 |
+
for i in range(len(images)):
|
49 |
+
# Save pages as images in the pdf
|
50 |
+
images[i].save('page'+ str(i) +'.jpg', 'JPEG')
|
51 |
+
x=str(i)
|
52 |
+
t='page'+x+'.jpg'
|
53 |
+
result.append(reader.readtext(t, detail = 0))
|
54 |
+
text = ""
|
55 |
+
for page in result:
|
56 |
+
page_text = " ".join(page)
|
57 |
+
text += page_text
|
58 |
+
model = pipeline("question-answering", model='deepset/roberta-base-squad2')
|
59 |
+
context = text
|
60 |
+
result = model(question=query, context=context)
|
61 |
+
print(f"Answer: {result['answer']}")
|
62 |
+
return result
|
63 |
+
|
64 |
+
convos = [] # list of all panel objects
|
65 |
+
|
66 |
+
def qa_result(_):
|
67 |
+
# saving pdf as a temp file
|
68 |
+
if file_input.value is not None:
|
69 |
+
file_input.save("/content/temp.pdf")
|
70 |
+
|
71 |
+
prompt_text = prompt.value
|
72 |
+
if prompt_text:
|
73 |
+
result = qa(file="/content/temp.pdf", query=prompt_text)
|
74 |
+
convos.extend([
|
75 |
+
pn.Row(
|
76 |
+
pn.panel("Q: ", width=10),
|
77 |
+
prompt_text,
|
78 |
+
width=600
|
79 |
+
),
|
80 |
+
pn.Row(
|
81 |
+
pn.panel("A: ", width=10),
|
82 |
+
pn.Column(
|
83 |
+
result["answer"],
|
84 |
+
)
|
85 |
+
)
|
86 |
+
])
|
87 |
+
#return convos
|
88 |
+
return pn.Column(*convos, margin=15, width=575, min_height=400)
|
89 |
+
|
90 |
+
qa_interactive = pn.panel(
|
91 |
+
pn.bind(qa_result, run_button),
|
92 |
+
loading_indicator=True,
|
93 |
+
)
|
94 |
+
|
95 |
+
output = pn.WidgetBox('*Output will show up here:*', qa_interactive, width=630, scroll=True)
|
96 |
+
|
97 |
+
# layout
|
98 |
+
pn.Column(
|
99 |
+
pn.pane.Markdown("""
|
100 |
+
Question Answering with your PDF file!
|
101 |
+
|
102 |
+
1) Upload a PDF. \n
|
103 |
+
2) Type a question and click "Run".
|
104 |
+
|
105 |
+
"""),
|
106 |
+
pn.Row(file_input),
|
107 |
+
output,
|
108 |
+
widgets
|
109 |
+
|
110 |
+
).servable()
|
requirements.txt
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
panel
|
2 |
+
transformers
|
3 |
+
pdf2image
|
4 |
+
easyocr
|