Spaces:
Runtime error
Runtime error
domenicrosati
commited on
Commit
β’
4c36cd4
1
Parent(s):
8890bde
add strict relevancy and scite badges and reranking
Browse files- README.md +0 -2
- app.py +94 -27
- requirements.txt +3 -0
README.md
CHANGED
@@ -9,5 +9,3 @@ app_file: app.py
|
|
9 |
pinned: false
|
10 |
license: cc-by-2.0
|
11 |
---
|
12 |
-
|
13 |
-
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
|
|
|
9 |
pinned: false
|
10 |
license: cc-by-2.0
|
11 |
---
|
|
|
|
app.py
CHANGED
@@ -2,15 +2,38 @@ import streamlit as st
|
|
2 |
from transformers import pipeline
|
3 |
import requests
|
4 |
from bs4 import BeautifulSoup
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
|
6 |
SCITE_API_KEY = st.secrets["SCITE_API_KEY"]
|
7 |
|
|
|
8 |
def remove_html(x):
|
9 |
soup = BeautifulSoup(x, 'html.parser')
|
10 |
text = soup.get_text()
|
11 |
return text
|
12 |
|
13 |
-
|
|
|
|
|
|
|
14 |
search = f"https://api.scite.ai/search?mode=citations&term={term}&limit={limit}&offset=0&user_slug=domenic-rosati-keW5&compute_aggregations=false"
|
15 |
req = requests.get(
|
16 |
search,
|
@@ -19,8 +42,9 @@ def search(term, limit=25):
|
|
19 |
}
|
20 |
)
|
21 |
return (
|
22 |
-
remove_html('\n'.join([
|
23 |
-
[(doc['doi'], doc['citations'], doc['title'])
|
|
|
24 |
)
|
25 |
|
26 |
|
@@ -39,25 +63,37 @@ def find_source(text, docs):
|
|
39 |
'source_title': doc[2],
|
40 |
'source_link': f"https://scite.ai/reports/{doc[0]}"
|
41 |
}
|
42 |
-
return
|
43 |
-
|
44 |
-
'text': text,
|
45 |
-
'from': '',
|
46 |
-
'supporting': '',
|
47 |
-
'source_title': '',
|
48 |
-
'source_link': ''
|
49 |
-
}
|
50 |
|
51 |
@st.experimental_singleton
|
52 |
def init_models():
|
53 |
-
|
54 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
55 |
|
56 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
57 |
|
58 |
|
59 |
-
|
60 |
-
|
|
|
61 |
<div class="container-fluid">
|
62 |
<div class="row align-items-start">
|
63 |
<div class="col-md-12 col-sm-12">
|
@@ -72,6 +108,22 @@ def card(title, context, score, link):
|
|
72 |
</div>
|
73 |
</div>
|
74 |
""", unsafe_allow_html=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
75 |
|
76 |
st.title("Scientific Question Answering with Citations")
|
77 |
|
@@ -85,8 +137,14 @@ st.markdown("""
|
|
85 |
""", unsafe_allow_html=True)
|
86 |
|
87 |
def run_query(query):
|
88 |
-
|
89 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
90 |
return st.markdown("""
|
91 |
<div class="container-fluid">
|
92 |
<div class="row align-items-start">
|
@@ -97,35 +155,44 @@ def run_query(query):
|
|
97 |
</div>
|
98 |
""", unsafe_allow_html=True)
|
99 |
|
|
|
|
|
|
|
|
|
|
|
|
|
100 |
results = []
|
101 |
model_results = qa_model(question=query, context=context, top_k=10)
|
102 |
for result in model_results:
|
103 |
support = find_source(result['answer'], orig_docs)
|
|
|
|
|
104 |
results.append({
|
105 |
"answer": support['text'],
|
106 |
"title": support['source_title'],
|
107 |
"link": support['source_link'],
|
108 |
"context": support['citation_statement'],
|
109 |
-
"score": result['score']
|
|
|
110 |
})
|
111 |
|
112 |
-
|
113 |
-
|
114 |
sorted_result = sorted(results, key=lambda x: x['score'], reverse=True)
|
115 |
sorted_result = list({
|
116 |
result['context']: result for result in sorted_result
|
117 |
}.values())
|
118 |
-
sorted_result = sorted(
|
119 |
-
|
120 |
|
121 |
for r in sorted_result:
|
122 |
answer = r["answer"]
|
123 |
-
ctx = remove_html(r["context"]).replace(answer, f"<mark>{answer}</mark>").replace(
|
124 |
-
|
|
|
125 |
score = round(r["score"], 4)
|
126 |
-
card(title, ctx, score, r['link'])
|
127 |
|
128 |
query = st.text_input("Ask scientific literature a question", "")
|
129 |
|
130 |
if query != "":
|
131 |
-
|
|
|
|
2 |
from transformers import pipeline
|
3 |
import requests
|
4 |
from bs4 import BeautifulSoup
|
5 |
+
from nltk.corpus import stopwords
|
6 |
+
import nltk
|
7 |
+
import string
|
8 |
+
from streamlit.components.v1 import html
|
9 |
+
from sentence_transformers.cross_encoder import CrossEncoder as CE
|
10 |
+
import numpy as np
|
11 |
+
from typing import List, Tuple
|
12 |
+
import torch
|
13 |
+
|
14 |
+
class CrossEncoder:
|
15 |
+
def __init__(self, model_path: str, **kwargs):
|
16 |
+
self.model = CE(model_path, **kwargs)
|
17 |
+
|
18 |
+
def predict(self, sentences: List[Tuple[str,str]], batch_size: int = 32, show_progress_bar: bool = True) -> List[float]:
|
19 |
+
return self.model.predict(
|
20 |
+
sentences=sentences,
|
21 |
+
batch_size=batch_size,
|
22 |
+
show_progress_bar=show_progress_bar)
|
23 |
+
|
24 |
|
25 |
SCITE_API_KEY = st.secrets["SCITE_API_KEY"]
|
26 |
|
27 |
+
|
28 |
def remove_html(x):
|
29 |
soup = BeautifulSoup(x, 'html.parser')
|
30 |
text = soup.get_text()
|
31 |
return text
|
32 |
|
33 |
+
|
34 |
+
def search(term, limit=10, clean=True, strict=True):
|
35 |
+
term = clean_query(term, clean=clean, strict=strict)
|
36 |
+
# heuristic, 2 searches strict and not? and then merge?
|
37 |
search = f"https://api.scite.ai/search?mode=citations&term={term}&limit={limit}&offset=0&user_slug=domenic-rosati-keW5&compute_aggregations=false"
|
38 |
req = requests.get(
|
39 |
search,
|
|
|
42 |
}
|
43 |
)
|
44 |
return (
|
45 |
+
[remove_html('\n'.join([cite['snippet'] for cite in doc['citations']])) for doc in req.json()['hits']],
|
46 |
+
[(doc['doi'], doc['citations'], doc['title'])
|
47 |
+
for doc in req.json()['hits']]
|
48 |
)
|
49 |
|
50 |
|
|
|
63 |
'source_title': doc[2],
|
64 |
'source_link': f"https://scite.ai/reports/{doc[0]}"
|
65 |
}
|
66 |
+
return None
|
67 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
68 |
|
69 |
@st.experimental_singleton
|
70 |
def init_models():
|
71 |
+
nltk.download('stopwords')
|
72 |
+
stop = set(stopwords.words('english') + list(string.punctuation))
|
73 |
+
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
74 |
+
question_answerer = pipeline(
|
75 |
+
"question-answering", model='sultan/BioM-ELECTRA-Large-SQuAD2-BioASQ8B',
|
76 |
+
device=device
|
77 |
+
)
|
78 |
+
reranker = CrossEncoder('cross-encoder/ms-marco-MiniLM-L-6-v2', device=device)
|
79 |
+
return question_answerer, reranker, stop, device
|
80 |
+
|
81 |
+
qa_model, reranker, stop, device = init_models()
|
82 |
|
83 |
+
def clean_query(query, strict=True, clean=True):
|
84 |
+
operator = ' '
|
85 |
+
if strict:
|
86 |
+
operator = ' AND '
|
87 |
+
query = operator.join(
|
88 |
+
[i for i in query.lower().split(' ') if clean and i not in stop])
|
89 |
+
if clean:
|
90 |
+
query = query.translate(str.maketrans('', '', string.punctuation))
|
91 |
+
return query
|
92 |
|
93 |
|
94 |
+
|
95 |
+
def card(title, context, score, link, supporting):
|
96 |
+
st.markdown(f"""
|
97 |
<div class="container-fluid">
|
98 |
<div class="row align-items-start">
|
99 |
<div class="col-md-12 col-sm-12">
|
|
|
108 |
</div>
|
109 |
</div>
|
110 |
""", unsafe_allow_html=True)
|
111 |
+
html(f"""
|
112 |
+
<div
|
113 |
+
class="scite-badge"
|
114 |
+
data-doi="{supporting}"
|
115 |
+
data-layout="horizontal"
|
116 |
+
data-show-zero="false"
|
117 |
+
data-show-labels="false"
|
118 |
+
data-tally-show="true"
|
119 |
+
/>
|
120 |
+
<script
|
121 |
+
async
|
122 |
+
type="application/javascript"
|
123 |
+
src="https://cdn.scite.ai/badge/scite-badge-latest.min.js">
|
124 |
+
</script>
|
125 |
+
""", width=None, height=42, scrolling=False)
|
126 |
+
|
127 |
|
128 |
st.title("Scientific Question Answering with Citations")
|
129 |
|
|
|
137 |
""", unsafe_allow_html=True)
|
138 |
|
139 |
def run_query(query):
|
140 |
+
if device == 'cpu':
|
141 |
+
limit = 50
|
142 |
+
context_limit = 10
|
143 |
+
else:
|
144 |
+
limit = 100
|
145 |
+
context_limit = 25
|
146 |
+
contexts, orig_docs = search(query, limit=limit)
|
147 |
+
if len(contexts) == 0 or not ''.join(contexts).strip():
|
148 |
return st.markdown("""
|
149 |
<div class="container-fluid">
|
150 |
<div class="row align-items-start">
|
|
|
155 |
</div>
|
156 |
""", unsafe_allow_html=True)
|
157 |
|
158 |
+
sentence_pairs = [[query, context] for context in contexts]
|
159 |
+
scores = reranker.predict(sentence_pairs, batch_size=limit, show_progress_bar=False)
|
160 |
+
hits = {contexts[idx]: scores[idx] for idx in range(len(scores))}
|
161 |
+
sorted_contexts = [k for k,v in sorted(hits.items(), key=lambda x: x[0], reverse=True)]
|
162 |
+
|
163 |
+
context = '\n'.join(sorted_contexts[:context_limit])
|
164 |
results = []
|
165 |
model_results = qa_model(question=query, context=context, top_k=10)
|
166 |
for result in model_results:
|
167 |
support = find_source(result['answer'], orig_docs)
|
168 |
+
if not support:
|
169 |
+
continue
|
170 |
results.append({
|
171 |
"answer": support['text'],
|
172 |
"title": support['source_title'],
|
173 |
"link": support['source_link'],
|
174 |
"context": support['citation_statement'],
|
175 |
+
"score": result['score'],
|
176 |
+
"doi": support["supporting"]
|
177 |
})
|
178 |
|
|
|
|
|
179 |
sorted_result = sorted(results, key=lambda x: x['score'], reverse=True)
|
180 |
sorted_result = list({
|
181 |
result['context']: result for result in sorted_result
|
182 |
}.values())
|
183 |
+
sorted_result = sorted(
|
184 |
+
sorted_result, key=lambda x: x['score'], reverse=True)
|
185 |
|
186 |
for r in sorted_result:
|
187 |
answer = r["answer"]
|
188 |
+
ctx = remove_html(r["context"]).replace(answer, f"<mark>{answer}</mark>").replace(
|
189 |
+
'<cite', '<a').replace('</cite', '</a').replace('data-doi="', 'href="https://scite.ai/reports/')
|
190 |
+
title = r.get("title", '').replace("_", " ")
|
191 |
score = round(r["score"], 4)
|
192 |
+
card(title, ctx, score, r['link'], r['doi'])
|
193 |
|
194 |
query = st.text_input("Ask scientific literature a question", "")
|
195 |
|
196 |
if query != "":
|
197 |
+
with st.spinner('Loading...'):
|
198 |
+
run_query(query)
|
requirements.txt
CHANGED
@@ -3,3 +3,6 @@ requests
|
|
3 |
beautifulsoup4
|
4 |
streamlit==1.2.0
|
5 |
torch
|
|
|
|
|
|
|
|
3 |
beautifulsoup4
|
4 |
streamlit==1.2.0
|
5 |
torch
|
6 |
+
nltk
|
7 |
+
sentence_transformers
|
8 |
+
numpy
|