File size: 14,551 Bytes
60b1427
 
 
65d5daa
60b1427
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b7bd5a2
 
 
 
1229bf2
60b1427
b7bd5a2
957d2c2
b7bd5a2
 
 
 
957d2c2
1bd8049
 
1229bf2
b7bd5a2
dde1577
b7bd5a2
957d2c2
 
 
b7bd5a2
 
 
1bd8049
 
 
957d2c2
 
dde1577
1bd8049
1229bf2
b7bd5a2
dde1577
b7bd5a2
957d2c2
 
 
b7bd5a2
 
 
1bd8049
 
 
957d2c2
1bd8049
b7bd5a2
1bd8049
 
 
 
60b1427
b7bd5a2
 
 
 
 
 
 
 
 
 
60b1427
 
 
 
 
 
b7bd5a2
60b1427
 
 
 
 
 
6cd4890
60b1427
 
 
 
 
d16f597
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
60b1427
 
6cd4890
 
 
b7bd5a2
 
 
 
 
d16f597
 
 
b7bd5a2
 
 
 
d16f597
60b1427
 
 
 
 
 
b7bd5a2
 
60b1427
 
b7bd5a2
 
60b1427
 
 
b7bd5a2
 
 
 
 
 
 
60b1427
 
 
2c2de78
 
 
60b1427
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b7bd5a2
 
 
 
60b1427
 
 
 
 
 
6cd4890
60b1427
b7bd5a2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1229bf2
b7bd5a2
 
 
 
1229bf2
b7bd5a2
 
 
 
1229bf2
b7bd5a2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
60b1427
b7bd5a2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1229bf2
b7bd5a2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
60b1427
b7bd5a2
 
 
 
 
 
 
60b1427
 
b7bd5a2
60b1427
b7bd5a2
 
1229bf2
b7bd5a2
60b1427
 
86deaaa
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
import streamlit as st
import pandas as pd
import torch
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
from peft import PeftModel
from text_processing import TextProcessor
import gc
from pathlib import Path

# Configure page
st.set_page_config(
    page_title="Biomedical Papers Analysis",
    page_icon="πŸ”¬",
    layout="wide"
)

# Initialize session state
if 'processed_data' not in st.session_state:
    st.session_state.processed_data = None
if 'summaries' not in st.session_state:
    st.session_state.summaries = None
if 'text_processor' not in st.session_state:
    st.session_state.text_processor = None
if 'processing_started' not in st.session_state:
    st.session_state.processing_started = False
if 'focused_summary_generated' not in st.session_state:
    st.session_state.focused_summary_generated = False

def load_model(model_type):
    """Load appropriate model based on type with proper memory management"""
    try:
        # Clear any existing cached data
        torch.cuda.empty_cache()
        gc.collect()
        
        if model_type == "summarize":
            base_model = AutoModelForSeq2SeqLM.from_pretrained(
                "facebook/bart-large-cnn",
                cache_dir="./models",
                low_cpu_mem_usage=True,
                torch_dtype=torch.float32
            )
            model = PeftModel.from_pretrained(
                base_model, 
                "pendar02/results",
                device_map="auto",
                torch_dtype=torch.float32
            )
            tokenizer = AutoTokenizer.from_pretrained(
                "facebook/bart-large-cnn",
                cache_dir="./models"
            )
        else:  # question_focused
            base_model = AutoModelForSeq2SeqLM.from_pretrained(
                "GanjinZero/biobart-base",
                cache_dir="./models",
                low_cpu_mem_usage=True,
                torch_dtype=torch.float32
            )
            model = PeftModel.from_pretrained(
                base_model, 
                "pendar02/biobart-finetune",
                device_map="auto",
                torch_dtype=torch.float32
            )
            tokenizer = AutoTokenizer.from_pretrained(
                "GanjinZero/biobart-base",
                cache_dir="./models"
            )
        
        model.eval()
        return model, tokenizer
    except Exception as e:
        st.error(f"Error loading model: {str(e)}")
        raise

def cleanup_model(model, tokenizer):
    """Properly cleanup model resources"""
    try:
        del model
        del tokenizer
        torch.cuda.empty_cache()
        gc.collect()
    except Exception:
        pass

@st.cache_data
def process_excel(uploaded_file):
    """Process uploaded Excel file"""
    try:
        df = pd.read_excel(uploaded_file)
        required_columns = ['Abstract', 'Article Title', 'Authors', 
                          'Source Title', 'Publication Year', 'DOI']
        
        # Check required columns
        missing_columns = [col for col in required_columns if col not in df.columns]
        if missing_columns:
            st.error(f"Missing required columns: {', '.join(missing_columns)}")
            return None
        
        return df[required_columns]
    except Exception as e:
        st.error(f"Error processing file: {str(e)}")
        return None

def preprocess_text(text):
    """Preprocess text to add appropriate formatting before summarization"""
    if not isinstance(text, str) or not text.strip():
        return text
        
    # Split text into sentences (basic implementation)
    sentences = [s.strip() for s in text.replace('. ', '.\n').split('\n')]
    
    # Remove empty sentences
    sentences = [s for s in sentences if s]
    
    # Join with proper line breaks
    formatted_text = '\n'.join(sentences)
    
    return formatted_text

def generate_summary(text, model, tokenizer):
    """Generate summary for single abstract"""
    if not isinstance(text, str) or not text.strip():
        return "No abstract available to summarize."
    
    # Check if abstract is too short
    word_count = len(text.split())
    if word_count < 50:  # Threshold for "short" abstracts
        return text  # Return original text for very short abstracts
    
    # Preprocess the text first
    formatted_text = preprocess_text(text)
    
    # Adjust generation parameters based on input length
    max_length = min(150, word_count + 50)  # Dynamic max length
    min_length = min(50, word_count)  # Dynamic min length
    
    inputs = tokenizer(formatted_text, return_tensors="pt", max_length=1024, truncation=True)
    
    with torch.no_grad():
        summary_ids = model.generate(
            **{
                "input_ids": inputs["input_ids"],
                "attention_mask": inputs["attention_mask"],
                "max_length": max_length,
                "min_length": min_length,
                "num_beams": 4,
                "length_penalty": 2.0,
                "early_stopping": True,
                "no_repeat_ngram_size": 3  # Prevent repetition of phrases
            }
        )
    
    summary = tokenizer.decode(summary_ids[0], skip_special_tokens=True)
    
    # Post-process summary
    if summary.lower() == text.lower() or len(summary.split()) / word_count > 0.9:
        return text  # Return original if summary is too similar
        
    return summary

def generate_focused_summary(question, abstracts, model, tokenizer):
    """Generate focused summary based on question"""
    # Preprocess each abstract
    formatted_abstracts = [preprocess_text(abstract) for abstract in abstracts]
    combined_input = f"Question: {question} Abstracts: " + " [SEP] ".join(formatted_abstracts)
    
    inputs = tokenizer(combined_input, return_tensors="pt", max_length=1024, truncation=True)
    
    with torch.no_grad():
        summary_ids = model.generate(
            **{
                "input_ids": inputs["input_ids"],
                "attention_mask": inputs["attention_mask"],
                "max_length": 200,
                "min_length": 50,
                "num_beams": 4,
                "length_penalty": 2.0,
                "early_stopping": True
            }
        )
    
    return tokenizer.decode(summary_ids[0], skip_special_tokens=True)

def main():
    st.title("πŸ”¬ Biomedical Papers Analysis")
    
    # File upload section
    uploaded_file = st.file_uploader(
        "Upload Excel file containing papers", 
        type=['xlsx', 'xls'],
        help="File must contain: Abstract, Article Title, Authors, Source Title, Publication Year, DOI"
    )
    
    # Question input - moved up but hidden initially
    question_container = st.empty()
    question = ""
    
    if uploaded_file is not None:
        # Process Excel file
        if st.session_state.processed_data is None:
            with st.spinner("Processing file..."):
                df = process_excel(uploaded_file)
                if df is not None:
                    st.session_state.processed_data = df.dropna(subset=["Abstract"])
        
        if st.session_state.processed_data is not None:
            df = st.session_state.processed_data
            st.write(f"πŸ“Š Loaded {len(df)} papers with abstracts")
            
            # Get question before processing
            with question_container:
                question = st.text_input(
                    "Enter your research question (optional):",
                    help="If provided, a question-focused summary will be generated after individual summaries"
                )
            
            # Single button for both processes
            if not st.session_state.get('processing_started', False):
                if st.button("Start Analysis"):
                    st.session_state.processing_started = True
            
            # Show processing status and results
            if st.session_state.get('processing_started', False):
                # Individual Summaries Section
                st.header("πŸ“ Individual Paper Summaries")
                
                if st.session_state.summaries is None:
                    try:
                        with st.spinner("Generating summaries..."):
                            # Load summarization model
                            model, tokenizer = load_model("summarize")
                            
                            # Process abstracts with real-time updates
                            summaries = []
                            progress_bar = st.progress(0)
                            summary_display = st.empty()
                            
                            for i, (_, row) in enumerate(df.iterrows()):
                                summary = generate_summary(row['Abstract'], model, tokenizer)
                                summaries.append(summary)
                                
                                # Update progress and show current summary
                                progress = (i + 1) / len(df)
                                progress_bar.progress(progress)
                                summary_display.write(f"Processing paper {i+1}/{len(df)}:\n{row['Article Title']}")
                                
                            st.session_state.summaries = summaries
                            
                            # Cleanup first model
                            cleanup_model(model, tokenizer)
                    
                    except Exception as e:
                        st.error(f"Error generating summaries: {str(e)}")
                
                # Display summaries with improved sorting
                if st.session_state.summaries is not None:
                    col1, col2 = st.columns(2)
                    with col1:
                        sort_options = ['Article Title', 'Authors', 'Publication Year', 'Source Title']
                        sort_column = st.selectbox("Sort by:", sort_options)
                    with col2:
                        ascending = st.checkbox("Ascending order", True)
                    
                    # Create display dataframe with formatted year
                    display_df = df.copy()
                    display_df['Summary'] = st.session_state.summaries
                    display_df['Publication Year'] = display_df['Publication Year'].astype(int)
                    sorted_df = display_df.sort_values(by=sort_column, ascending=ascending)
                    
                    # Apply custom formatting
                    st.markdown("""
                    <style>
                    .stDataFrame {
                        font-size: 16px;
                    }
                    .stDataFrame td {
                        white-space: normal !important;
                        padding: 8px !important;
                    }
                    </style>
                    """, unsafe_allow_html=True)
                    
                    st.dataframe(
                        sorted_df[['Article Title', 'Authors', 'Source Title', 
                                 'Publication Year', 'DOI', 'Summary']],
                        hide_index=True
                    )
                
                # Question-focused Summary Section (only if question provided)
                if question.strip():
                    st.header("❓ Question-focused Summary")
                    
                    if not st.session_state.get('focused_summary_generated', False):
                        try:
                            with st.spinner("Analyzing relevant papers..."):
                                # Initialize text processor if needed
                                if st.session_state.text_processor is None:
                                    st.session_state.text_processor = TextProcessor()
                                
                                # Find relevant abstracts
                                results = st.session_state.text_processor.find_most_relevant_abstracts(
                                    question,
                                    df['Abstract'].tolist(),
                                    top_k=5
                                )
                                
                                # Load question-focused model
                                model, tokenizer = load_model("question_focused")
                                
                                # Generate focused summary
                                relevant_abstracts = df['Abstract'].iloc[results['top_indices']].tolist()
                                focused_summary = generate_focused_summary(
                                    question,
                                    relevant_abstracts,
                                    model,
                                    tokenizer
                                )
                                
                                # Store results
                                st.session_state.focused_summary = focused_summary
                                st.session_state.relevant_papers = df.iloc[results['top_indices']]
                                st.session_state.relevance_scores = results['scores']
                                st.session_state.focused_summary_generated = True
                                
                                # Cleanup second model
                                cleanup_model(model, tokenizer)
                        
                        except Exception as e:
                            st.error(f"Error generating focused summary: {str(e)}")
                    
                    # Display focused summary results
                    if st.session_state.get('focused_summary_generated', False):
                        st.subheader("Summary")
                        st.write(st.session_state.focused_summary)
                        
                        st.subheader("Most Relevant Papers")
                        relevant_papers = st.session_state.relevant_papers[
                            ['Article Title', 'Authors', 'Publication Year', 'DOI']
                        ].copy()
                        relevant_papers['Relevance Score'] = st.session_state.relevance_scores
                        relevant_papers['Publication Year'] = relevant_papers['Publication Year'].astype(int)
                        st.dataframe(relevant_papers, hide_index=True)

if __name__ == "__main__":
    main()