Subahdip627 commited on
Commit
57718f7
Β·
verified Β·
1 Parent(s): b7926fd

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +451 -178
app.py CHANGED
@@ -1,6 +1,6 @@
1
  """
2
  PaperSearchPro - STRICT Hierarchical Keyword Search
3
- INCLUDES FULL ABSTRACT in results
4
  """
5
 
6
  import requests
@@ -10,6 +10,264 @@ from datetime import datetime
10
  import gradio as gr
11
  import tempfile
12
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
13
  class PaperSearchMobile:
14
  def __init__(self):
15
  self.base_url = "https://api.openalex.org/works"
@@ -45,10 +303,6 @@ class PaperSearchMobile:
45
  page = 1
46
  max_attempts = 50
47
 
48
- print(f"\nπŸ” Searching: {query}")
49
- print(f"πŸ“… Year range: {year_from} - {year_to}")
50
- print(f"πŸ“Š Max papers: {max_papers}")
51
-
52
  # Fetch papers
53
  while len(all_papers) < max_papers and page <= max_attempts:
54
  params = {
@@ -67,11 +321,8 @@ class PaperSearchMobile:
67
  results = data.get("results", [])
68
 
69
  if not results:
70
- print(f" ⏹️ No more results at page {page}")
71
  break
72
 
73
- print(f" πŸ“₯ Page {page}: Got {len(results)} papers")
74
-
75
  for work in results:
76
  paper = self._extract_paper_info(work)
77
  all_papers.append(paper)
@@ -80,23 +331,15 @@ class PaperSearchMobile:
80
  time.sleep(0.3)
81
 
82
  except Exception as e:
83
- print(f" ❌ Error on page {page}: {e}")
84
  break
85
 
86
- print(f"\nβœ… Fetched {len(all_papers)} papers total")
87
-
88
- # ========== STRICT HIERARCHICAL FILTERING ==========
89
  keyword_list = []
90
  if keywords and keywords.strip():
91
  keyword_list = [k.strip().lower() for k in keywords.split(',') if k.strip()]
92
 
93
  if len(keyword_list) >= 2:
94
- print(f"\nπŸ”‘ STRICT Filtering: MUST contain BOTH '{keyword_list[0]}' AND '{keyword_list[1]}'")
95
-
96
  scored_papers = []
97
- papers_missing = 0
98
- papers_with_abstract = 0
99
-
100
  for paper in all_papers:
101
  title_text = paper['Title'].lower()
102
  abstract_text = paper.get('Abstract', '').lower()
@@ -109,12 +352,10 @@ class PaperSearchMobile:
109
  kw2_present = keyword_list[1] in combined_text
110
 
111
  if not (kw1_present and kw2_present):
112
- papers_missing += 1
113
  continue
114
 
115
  # Check abstract
116
  if (keyword_list[0] in abstract_text) or (keyword_list[1] in abstract_text):
117
- papers_with_abstract += 1
118
  paper['abstract_match'] = "Yes"
119
  else:
120
  paper['abstract_match'] = "No"
@@ -131,9 +372,7 @@ class PaperSearchMobile:
131
  if paper['abstract_match'] == "Yes":
132
  score += 1
133
 
134
- # Truncate abstract for display (but keep full for export)
135
  paper['Abstract_Short'] = paper['Abstract'][:200] + "..." if len(paper['Abstract']) > 200 else paper['Abstract']
136
-
137
  paper['score'] = score
138
  paper['matched'] = ', '.join(matched)
139
  paper['required_matched'] = f"βœ“ {keyword_list[0]}, βœ“ {keyword_list[1]}"
@@ -142,10 +381,6 @@ class PaperSearchMobile:
142
 
143
  scored_papers.sort(key=lambda x: x['score'], reverse=True)
144
  all_papers = scored_papers
145
-
146
- print(f" βœ… Found {len(all_papers)} papers with BOTH required keywords")
147
- print(f" πŸ“„ Papers with keywords in abstract: {papers_with_abstract}")
148
- print(f" ❌ Papers filtered out: {papers_missing}")
149
 
150
  # Prepare display
151
  if all_papers:
@@ -153,35 +388,21 @@ class PaperSearchMobile:
153
  p['Rank'] = i + 1
154
 
155
  df = pd.DataFrame(all_papers)
156
-
157
- # Display columns (with truncated abstract)
158
  display_cols = ["Rank", "score", "Title", "Authors", "Year", "Citations", "Journal", "abstract_match", "Abstract_Short"]
159
-
160
  display_df = df[display_cols]
161
 
162
  rename_dict = {
163
  "score": "Relevance",
164
  "abstract_match": "In Abstract",
165
- "Abstract_Short": "Abstract (Preview)"
166
  }
167
-
168
  display_df = display_df.rename(columns=rename_dict)
169
 
170
- # Create summary
171
  years_list = [p['Year'] for p in all_papers if p['Year'] != 'N/A']
172
 
173
- summary = f"""
174
- ### πŸ“Š Search Results
175
- - **{len(all_papers)} papers** found (contain BOTH '{keyword_list[0]}' AND '{keyword_list[1]}')
176
- - **{df['abstract_match'].value_counts().get('Yes', 0)} papers** have keywords in abstract
177
- - **Average relevance score:** {df['score'].mean():.1f}
178
- - **Year range:** {min(years_list)} - {max(years_list)}
179
- - **Sort order:** {sort_by} ({sort_order})
180
- """
181
-
182
- return display_df, summary, all_papers
183
 
184
- return pd.DataFrame(), "### πŸ“Š No papers found matching your criteria", []
185
 
186
  def _extract_paper_info(self, work):
187
  def get_authors(authorships):
@@ -194,7 +415,7 @@ class PaperSearchMobile:
194
  authors.append(name)
195
  return ", ".join(authors) + ("..." if len(authorships) > 3 else "")
196
 
197
- # Get FULL abstract (reconstructed from inverted index)
198
  abstract = ""
199
  if work.get("abstract_inverted_index"):
200
  words = []
@@ -214,30 +435,21 @@ class PaperSearchMobile:
214
  "Journal": (work.get("primary_location", {}).get("source") or {}).get("display_name", "Unknown"),
215
  "DOI": work.get("doi", "N/A"),
216
  "URL": work.get("id", "N/A"),
217
- "Abstract": abstract, # Full abstract stored for export
218
  "matched": ""
219
  }
220
 
221
- # ============================================
222
- # EXPORT FUNCTION - INCLUDES FULL ABSTRACT
223
- # ============================================
224
  def download_results(papers, fname, fmt):
225
- """Create downloadable file with FULL abstract included"""
226
  if not papers:
227
  return None, "No papers to export"
228
 
229
- # Create DataFrame with ALL fields including full abstract
230
  df = pd.DataFrame(papers)
231
-
232
- # Reorder columns for better readability
233
  column_order = ['Rank', 'score', 'Title', 'Authors', 'Year', 'Citations', 'Journal',
234
  'DOI', 'URL', 'Abstract', 'matched', 'required_matched', 'abstract_match']
235
 
236
- # Only include columns that exist
237
  available_cols = [col for col in column_order if col in df.columns]
238
  df = df[available_cols]
239
 
240
- # Rename columns for export
241
  df = df.rename(columns={
242
  'score': 'Relevance_Score',
243
  'matched': 'All_Keywords_Found',
@@ -251,154 +463,227 @@ def download_results(papers, fname, fmt):
251
  if fmt == "CSV":
252
  filename = f"{fname}_with_abstracts_{timestamp}.csv"
253
  df.to_csv(tmp.name, index=False)
254
- else: # Excel
255
  filename = f"{fname}_with_abstracts_{timestamp}.xlsx"
256
-
257
- # Create Excel with multiple sheets
258
  with pd.ExcelWriter(tmp.name, engine='openpyxl') as writer:
259
- # Main results sheet
260
  df.to_excel(writer, sheet_name='Results', index=False)
261
 
262
- # Summary sheet
263
  summary_data = {
264
- 'Metric': ['Total Papers', 'Date Exported', 'Search Query', 'Keywords Used'],
265
- 'Value': [
266
- len(papers),
267
- datetime.now().strftime('%Y-%m-%d %H:%M:%S'),
268
- 'User query',
269
- ', '.join(set([k for p in papers for k in p.get('matched', '').split(', ') if k]))
270
- ]
271
  }
272
  pd.DataFrame(summary_data).to_excel(writer, sheet_name='Summary', index=False)
273
 
274
- return tmp.name, f"βœ… Ready to download: {filename} (includes FULL abstracts)"
275
 
276
  # ============================================
277
- # GRADIO INTERFACE
278
  # ============================================
279
  def create_interface():
280
- with gr.Blocks(title="PaperSearchPro", theme=gr.themes.Soft()) as demo:
281
- gr.Markdown("""
282
- # πŸ“± PaperSearchPro - STRICT Hierarchical Search
283
- ### Both FIRST and SECOND keywords are REQUIRED
284
-
285
- <div style='background-color: #e8f4fd; padding: 15px; border-radius: 10px; margin: 10px 0;'>
286
- <h3>πŸ” Search Scope:</h3>
287
- <p>βœ… <strong>Searching in:</strong> TITLE + ABSTRACT + AUTHORS</p>
288
- <p>πŸ“„ <strong>Abstracts:</strong> FULL abstracts included in downloadable files</p>
289
  </div>
290
  """)
291
 
292
- with gr.Column():
293
- query = gr.Textbox(
294
- label="πŸ” Search Topic",
295
- value="photoacoustic imaging"
296
- )
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
297
 
298
  keywords = gr.Textbox(
299
- label="πŸ”‘ Keywords (comma-separated)",
300
- value="blood, photoacoustics",
301
- info="First two keywords are REQUIRED"
 
302
  )
303
 
 
304
  with gr.Row():
305
- sort_by = gr.Dropdown(
306
- label="Sort By",
307
- choices=["relevance", "citations", "year"],
308
- value="relevance"
309
- )
310
- sort_order = gr.Dropdown(
311
- label="Order",
312
- choices=["descending", "ascending"],
313
- value="descending"
314
- )
315
-
316
- with gr.Row():
317
- max_papers = gr.Slider(
318
- label="Max Papers to Fetch",
319
- minimum=100,
320
- maximum=10000,
321
- value=2000,
322
- step=100
323
- )
324
- open_access = gr.Checkbox(label="Open Access Only")
 
 
325
 
 
326
  with gr.Row():
327
- year_from = gr.Number(
328
- label="Year From",
329
- value=2000,
330
- minimum=1900,
331
- maximum=2026
332
- )
333
- year_to = gr.Number(
334
- label="Year To",
335
- value=2026,
336
- minimum=1900,
337
- maximum=2026
338
- )
 
 
 
 
 
339
 
340
- search_btn = gr.Button("πŸ” SEARCH", variant="primary", size="lg")
 
 
 
 
 
 
341
 
342
- summary = gr.Markdown("Ready")
 
 
 
 
 
 
 
 
 
 
 
 
343
 
344
- # Results with abstract preview
345
  results = gr.Dataframe(
346
- label="Results (with Abstract Preview)",
347
- wrap=True,
348
- row_count=15
 
349
  )
350
  papers_state = gr.State([])
351
 
352
- # Export section with clear indication of abstracts
353
- gr.Markdown("---")
354
- gr.Markdown("### πŸ’Ύ Download Results (FULL ABSTRACTS INCLUDED)")
355
- with gr.Row():
356
- filename = gr.Textbox(
357
- label="Filename",
358
- value="papers_results",
359
- placeholder="Enter filename without extension"
360
- )
361
- format_type = gr.Dropdown(
362
- label="Format",
363
- choices=["CSV", "Excel"],
364
- value="CSV"
365
- )
366
-
367
  with gr.Row():
368
- export_btn = gr.Button("πŸ“₯ DOWNLOAD (with Full Abstracts)", variant="secondary")
369
- export_status = gr.Textbox(label="", visible=True)
370
- download_file = gr.File(label="Click to Download", visible=False)
371
-
372
- searcher = PaperSearchMobile()
373
-
374
- def search_wrapper(*args):
375
- df, summary_text, papers = searcher.search_papers(*args)
376
- return df, summary_text, papers
 
 
 
 
 
 
377
 
378
- search_btn.click(
379
- search_wrapper,
380
- inputs=[query, keywords, sort_by, sort_order, max_papers, year_from, year_to, open_access],
381
- outputs=[results, summary, papers_state]
382
  )
 
 
383
 
384
- def export_wrapper(papers, fname, fmt):
385
- file_path, message = download_results(papers, fname, fmt)
386
- return file_path, message, gr.File(visible=True) if file_path else gr.File(visible=False)
387
-
388
- export_btn.click(
389
- export_wrapper,
390
- inputs=[papers_state, filename, format_type],
391
- outputs=[download_file, export_status, download_file]
 
 
 
 
 
 
 
392
  )
393
 
394
- # Information about export
395
- gr.Markdown("""
396
- ---
397
- ### πŸ“‹ Export Information:
398
- - **CSV/Excel files include FULL abstracts** (not truncated)
399
- - **Columns included:** Title, Authors, Year, Citations, Journal, DOI, URL, FULL Abstract, Keywords matched
400
- - **Excel format** includes a Summary sheet with metadata
401
- """)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
402
 
403
  return demo
404
 
@@ -407,16 +692,4 @@ def create_interface():
407
  # ============================================
408
  if __name__ == "__main__":
409
  demo = create_interface()
410
-
411
- print("\n" + "="*90)
412
- print("πŸš€ PaperSearchPro - WITH FULL ABSTRACTS IN EXPORT")
413
- print("="*90)
414
- print("\nβœ… NEW FEATURES:")
415
- print(" β€’ FULL abstracts included in CSV/Excel downloads")
416
- print(" β€’ Abstract preview shown in results table")
417
- print(" β€’ Excel format includes Summary sheet")
418
- print(" β€’ All paper metadata preserved")
419
- print("\n🌐 Deployed on Hugging Face Spaces - permanent link!")
420
- print("="*90 + "\n")
421
-
422
  demo.launch()
 
1
  """
2
  PaperSearchPro - STRICT Hierarchical Keyword Search
3
+ PREMIUM GUI with White Background and Modern Design
4
  """
5
 
6
  import requests
 
10
  import gradio as gr
11
  import tempfile
12
 
13
+ # Custom CSS for premium white background design
14
+ custom_css = """
15
+ :root {
16
+ --primary-color: #2563eb;
17
+ --secondary-color: #7c3aed;
18
+ --success-color: #10b981;
19
+ --warning-color: #f59e0b;
20
+ --danger-color: #ef4444;
21
+ --text-primary: #1f2937;
22
+ --text-secondary: #6b7280;
23
+ --border-color: #e5e7eb;
24
+ --bg-white: #ffffff;
25
+ --bg-light: #f9fafb;
26
+ --shadow-sm: 0 1px 2px 0 rgba(0, 0, 0, 0.05);
27
+ --shadow-md: 0 4px 6px -1px rgba(0, 0, 0, 0.1);
28
+ --shadow-lg: 0 10px 15px -3px rgba(0, 0, 0, 0.1);
29
+ --radius-sm: 0.375rem;
30
+ --radius-md: 0.5rem;
31
+ --radius-lg: 0.75rem;
32
+ }
33
+
34
+ /* Main container */
35
+ .gradio-container {
36
+ background-color: var(--bg-light) !important;
37
+ font-family: 'Inter', -apple-system, BlinkMacSystemFont, sans-serif;
38
+ }
39
+
40
+ /* White cards */
41
+ .white-card {
42
+ background-color: var(--bg-white);
43
+ border-radius: var(--radius-lg);
44
+ padding: 1.5rem;
45
+ box-shadow: var(--shadow-sm);
46
+ border: 1px solid var(--border-color);
47
+ transition: box-shadow 0.3s ease;
48
+ }
49
+
50
+ .white-card:hover {
51
+ box-shadow: var(--shadow-md);
52
+ }
53
+
54
+ /* Headers */
55
+ h1, h2, h3 {
56
+ color: var(--text-primary) !important;
57
+ font-weight: 600 !important;
58
+ }
59
+
60
+ h1 {
61
+ font-size: 2rem !important;
62
+ margin-bottom: 1rem !important;
63
+ }
64
+
65
+ h1 span {
66
+ background: linear-gradient(135deg, var(--primary-color), var(--secondary-color));
67
+ -webkit-background-clip: text;
68
+ -webkit-text-fill-color: transparent;
69
+ }
70
+
71
+ /* Input fields */
72
+ .gr-box, .gr-input, .gr-textarea {
73
+ background-color: var(--bg-white) !important;
74
+ border: 1px solid var(--border-color) !important;
75
+ border-radius: var(--radius-md) !important;
76
+ padding: 0.5rem !important;
77
+ transition: border-color 0.3s, box-shadow 0.3s !important;
78
+ }
79
+
80
+ .gr-box:focus, .gr-input:focus, .gr-textarea:focus {
81
+ border-color: var(--primary-color) !important;
82
+ box-shadow: 0 0 0 3px rgba(37, 99, 235, 0.1) !important;
83
+ outline: none !important;
84
+ }
85
+
86
+ /* Buttons */
87
+ .gr-button-primary {
88
+ background: linear-gradient(135deg, var(--primary-color), var(--secondary-color)) !important;
89
+ border: none !important;
90
+ color: white !important;
91
+ font-weight: 600 !important;
92
+ padding: 0.75rem 1.5rem !important;
93
+ border-radius: var(--radius-md) !important;
94
+ transition: transform 0.2s, box-shadow 0.2s !important;
95
+ }
96
+
97
+ .gr-button-primary:hover {
98
+ transform: translateY(-2px);
99
+ box-shadow: var(--shadow-md) !important;
100
+ }
101
+
102
+ .gr-button-primary:active {
103
+ transform: translateY(0);
104
+ }
105
+
106
+ .gr-button-secondary {
107
+ background-color: var(--bg-white) !important;
108
+ border: 1px solid var(--border-color) !important;
109
+ color: var(--text-primary) !important;
110
+ font-weight: 500 !important;
111
+ padding: 0.75rem 1.5rem !important;
112
+ border-radius: var(--radius-md) !important;
113
+ transition: all 0.2s !important;
114
+ }
115
+
116
+ .gr-button-secondary:hover {
117
+ background-color: var(--bg-light) !important;
118
+ border-color: var(--text-secondary) !important;
119
+ }
120
+
121
+ /* Labels and text */
122
+ label {
123
+ color: var(--text-secondary) !important;
124
+ font-size: 0.875rem !important;
125
+ font-weight: 500 !important;
126
+ margin-bottom: 0.25rem !important;
127
+ }
128
+
129
+ /* Dataframe/Table */
130
+ .gr-dataframe {
131
+ background-color: var(--bg-white) !important;
132
+ border: 1px solid var(--border-color) !important;
133
+ border-radius: var(--radius-lg) !important;
134
+ overflow: hidden !important;
135
+ }
136
+
137
+ .gr-dataframe table {
138
+ border-collapse: separate !important;
139
+ border-spacing: 0 !important;
140
+ }
141
+
142
+ .gr-dataframe th {
143
+ background-color: var(--bg-light) !important;
144
+ color: var(--text-primary) !important;
145
+ font-weight: 600 !important;
146
+ font-size: 0.875rem !important;
147
+ padding: 0.75rem !important;
148
+ border-bottom: 2px solid var(--border-color) !important;
149
+ }
150
+
151
+ .gr-dataframe td {
152
+ padding: 0.75rem !important;
153
+ border-bottom: 1px solid var(--border-color) !important;
154
+ color: var(--text-primary) !important;
155
+ }
156
+
157
+ .gr-dataframe tr:last-child td {
158
+ border-bottom: none !important;
159
+ }
160
+
161
+ .gr-dataframe tr:hover td {
162
+ background-color: var(--bg-light) !important;
163
+ }
164
+
165
+ /* Slider */
166
+ .gr-slider {
167
+ background: linear-gradient(90deg, var(--primary-color), var(--secondary-color)) !important;
168
+ }
169
+
170
+ /* Checkbox */
171
+ .gr-checkbox {
172
+ accent-color: var(--primary-color) !important;
173
+ }
174
+
175
+ /* Tabs */
176
+ .gr-tabs {
177
+ background-color: var(--bg-white) !important;
178
+ border-radius: var(--radius-lg) !important;
179
+ padding: 0.5rem !important;
180
+ }
181
+
182
+ .gr-tabs .tab-nav {
183
+ border-bottom: 2px solid var(--border-color) !important;
184
+ }
185
+
186
+ .gr-tabs .tab-nav button {
187
+ color: var(--text-secondary) !important;
188
+ font-weight: 500 !important;
189
+ padding: 0.75rem 1.5rem !important;
190
+ border: none !important;
191
+ background: transparent !important;
192
+ transition: all 0.2s !important;
193
+ }
194
+
195
+ .gr-tabs .tab-nav button.selected {
196
+ color: var(--primary-color) !important;
197
+ border-bottom: 2px solid var(--primary-color) !important;
198
+ }
199
+
200
+ /* Info boxes */
201
+ .info-box {
202
+ background: linear-gradient(135deg, #f0f9ff, #e6f7ff);
203
+ border-left: 4px solid var(--primary-color);
204
+ border-radius: var(--radius-md);
205
+ padding: 1rem;
206
+ margin: 1rem 0;
207
+ }
208
+
209
+ .warning-box {
210
+ background: linear-gradient(135deg, #fef3c7, #fffbeb);
211
+ border-left: 4px solid var(--warning-color);
212
+ border-radius: var(--radius-md);
213
+ padding: 1rem;
214
+ margin: 1rem 0;
215
+ }
216
+
217
+ .success-box {
218
+ background: linear-gradient(135deg, #d1fae5, #ecfdf5);
219
+ border-left: 4px solid var(--success-color);
220
+ border-radius: var(--radius-md);
221
+ padding: 1rem;
222
+ margin: 1rem 0;
223
+ }
224
+
225
+ /* Keyword badges */
226
+ .keyword-badge {
227
+ display: inline-block;
228
+ padding: 0.25rem 0.75rem;
229
+ border-radius: 9999px;
230
+ font-size: 0.875rem;
231
+ font-weight: 500;
232
+ margin: 0.25rem;
233
+ }
234
+
235
+ .keyword-required {
236
+ background-color: #fee2e2;
237
+ color: var(--danger-color);
238
+ border: 1px solid #fecaca;
239
+ }
240
+
241
+ .keyword-optional {
242
+ background-color: #dcfce7;
243
+ color: var(--success-color);
244
+ border: 1px solid #bbf7d0;
245
+ }
246
+
247
+ /* Metrics cards */
248
+ .metric-card {
249
+ background-color: var(--bg-white);
250
+ border-radius: var(--radius-lg);
251
+ padding: 1.25rem;
252
+ box-shadow: var(--shadow-sm);
253
+ border: 1px solid var(--border-color);
254
+ text-align: center;
255
+ }
256
+
257
+ .metric-value {
258
+ font-size: 1.875rem;
259
+ font-weight: 700;
260
+ color: var(--primary-color);
261
+ line-height: 1;
262
+ }
263
+
264
+ .metric-label {
265
+ font-size: 0.875rem;
266
+ color: var(--text-secondary);
267
+ margin-top: 0.5rem;
268
+ }
269
+ """
270
+
271
  class PaperSearchMobile:
272
  def __init__(self):
273
  self.base_url = "https://api.openalex.org/works"
 
303
  page = 1
304
  max_attempts = 50
305
 
 
 
 
 
306
  # Fetch papers
307
  while len(all_papers) < max_papers and page <= max_attempts:
308
  params = {
 
321
  results = data.get("results", [])
322
 
323
  if not results:
 
324
  break
325
 
 
 
326
  for work in results:
327
  paper = self._extract_paper_info(work)
328
  all_papers.append(paper)
 
331
  time.sleep(0.3)
332
 
333
  except Exception as e:
 
334
  break
335
 
336
+ # Strict hierarchical filtering
 
 
337
  keyword_list = []
338
  if keywords and keywords.strip():
339
  keyword_list = [k.strip().lower() for k in keywords.split(',') if k.strip()]
340
 
341
  if len(keyword_list) >= 2:
 
 
342
  scored_papers = []
 
 
 
343
  for paper in all_papers:
344
  title_text = paper['Title'].lower()
345
  abstract_text = paper.get('Abstract', '').lower()
 
352
  kw2_present = keyword_list[1] in combined_text
353
 
354
  if not (kw1_present and kw2_present):
 
355
  continue
356
 
357
  # Check abstract
358
  if (keyword_list[0] in abstract_text) or (keyword_list[1] in abstract_text):
 
359
  paper['abstract_match'] = "Yes"
360
  else:
361
  paper['abstract_match'] = "No"
 
372
  if paper['abstract_match'] == "Yes":
373
  score += 1
374
 
 
375
  paper['Abstract_Short'] = paper['Abstract'][:200] + "..." if len(paper['Abstract']) > 200 else paper['Abstract']
 
376
  paper['score'] = score
377
  paper['matched'] = ', '.join(matched)
378
  paper['required_matched'] = f"βœ“ {keyword_list[0]}, βœ“ {keyword_list[1]}"
 
381
 
382
  scored_papers.sort(key=lambda x: x['score'], reverse=True)
383
  all_papers = scored_papers
 
 
 
 
384
 
385
  # Prepare display
386
  if all_papers:
 
388
  p['Rank'] = i + 1
389
 
390
  df = pd.DataFrame(all_papers)
 
 
391
  display_cols = ["Rank", "score", "Title", "Authors", "Year", "Citations", "Journal", "abstract_match", "Abstract_Short"]
 
392
  display_df = df[display_cols]
393
 
394
  rename_dict = {
395
  "score": "Relevance",
396
  "abstract_match": "In Abstract",
397
+ "Abstract_Short": "Abstract Preview"
398
  }
 
399
  display_df = display_df.rename(columns=rename_dict)
400
 
 
401
  years_list = [p['Year'] for p in all_papers if p['Year'] != 'N/A']
402
 
403
+ return display_df, all_papers, years_list, keyword_list
 
 
 
 
 
 
 
 
 
404
 
405
+ return pd.DataFrame(), [], [], []
406
 
407
  def _extract_paper_info(self, work):
408
  def get_authors(authorships):
 
415
  authors.append(name)
416
  return ", ".join(authors) + ("..." if len(authorships) > 3 else "")
417
 
418
+ # Get FULL abstract
419
  abstract = ""
420
  if work.get("abstract_inverted_index"):
421
  words = []
 
435
  "Journal": (work.get("primary_location", {}).get("source") or {}).get("display_name", "Unknown"),
436
  "DOI": work.get("doi", "N/A"),
437
  "URL": work.get("id", "N/A"),
438
+ "Abstract": abstract,
439
  "matched": ""
440
  }
441
 
 
 
 
442
  def download_results(papers, fname, fmt):
 
443
  if not papers:
444
  return None, "No papers to export"
445
 
 
446
  df = pd.DataFrame(papers)
 
 
447
  column_order = ['Rank', 'score', 'Title', 'Authors', 'Year', 'Citations', 'Journal',
448
  'DOI', 'URL', 'Abstract', 'matched', 'required_matched', 'abstract_match']
449
 
 
450
  available_cols = [col for col in column_order if col in df.columns]
451
  df = df[available_cols]
452
 
 
453
  df = df.rename(columns={
454
  'score': 'Relevance_Score',
455
  'matched': 'All_Keywords_Found',
 
463
  if fmt == "CSV":
464
  filename = f"{fname}_with_abstracts_{timestamp}.csv"
465
  df.to_csv(tmp.name, index=False)
466
+ else:
467
  filename = f"{fname}_with_abstracts_{timestamp}.xlsx"
 
 
468
  with pd.ExcelWriter(tmp.name, engine='openpyxl') as writer:
 
469
  df.to_excel(writer, sheet_name='Results', index=False)
470
 
 
471
  summary_data = {
472
+ 'Metric': ['Total Papers', 'Date Exported', 'Export Format'],
473
+ 'Value': [len(papers), datetime.now().strftime('%Y-%m-%d %H:%M:%S'), fmt]
 
 
 
 
 
474
  }
475
  pd.DataFrame(summary_data).to_excel(writer, sheet_name='Summary', index=False)
476
 
477
+ return tmp.name, f"βœ… Ready to download: {filename}"
478
 
479
  # ============================================
480
+ # PREMIUM GRADIO INTERFACE
481
  # ============================================
482
  def create_interface():
483
+ with gr.Blocks(title="PaperSearchPro", theme=gr.themes.Soft(), css=custom_css) as demo:
484
+ # Header with gradient title
485
+ gr.HTML("""
486
+ <div style='text-align: center; padding: 2rem 0;'>
487
+ <h1><span>πŸ“± PaperSearchPro</span></h1>
488
+ <p style='color: #6b7280; font-size: 1.1rem;'>Academic Paper Search with Hierarchical Keywords</p>
 
 
 
489
  </div>
490
  """)
491
 
492
+ # Main content in white card
493
+ with gr.Column(elem_classes="white-card"):
494
+ # Search section
495
+ with gr.Row():
496
+ with gr.Column(scale=2):
497
+ query = gr.Textbox(
498
+ label="πŸ” Search Topic",
499
+ placeholder="Enter your research topic...",
500
+ value="photoacoustic imaging",
501
+ elem_classes="gr-input"
502
+ )
503
+
504
+ with gr.Column(scale=1):
505
+ max_papers = gr.Slider(
506
+ label="πŸ“Š Max Papers",
507
+ minimum=100,
508
+ maximum=10000,
509
+ value=2000,
510
+ step=100,
511
+ elem_classes="gr-slider"
512
+ )
513
+
514
+ # Keywords section with badges
515
+ gr.HTML("""
516
+ <div style='margin: 1rem 0;'>
517
+ <h3 style='margin-bottom: 0.5rem;'>πŸ”‘ Keyword Hierarchy</h3>
518
+ <div>
519
+ <span class='keyword-badge keyword-required'>Keyword 1: REQUIRED</span>
520
+ <span class='keyword-badge keyword-required'>Keyword 2: REQUIRED</span>
521
+ <span class='keyword-badge keyword-optional'>Keyword 3+: Optional</span>
522
+ </div>
523
+ </div>
524
+ """)
525
 
526
  keywords = gr.Textbox(
527
+ label="Enter keywords (comma-separated)",
528
+ placeholder="blood, photoacoustics, imaging, cancer",
529
+ value="blood, photoacoustics, imaging",
530
+ elem_classes="gr-input"
531
  )
532
 
533
+ # Filters in a grid
534
  with gr.Row():
535
+ with gr.Column():
536
+ sort_by = gr.Dropdown(
537
+ label="Sort By",
538
+ choices=["relevance", "citations", "year"],
539
+ value="relevance",
540
+ elem_classes="gr-input"
541
+ )
542
+
543
+ with gr.Column():
544
+ sort_order = gr.Dropdown(
545
+ label="Order",
546
+ choices=["descending", "ascending"],
547
+ value="descending",
548
+ elem_classes="gr-input"
549
+ )
550
+
551
+ with gr.Column():
552
+ open_access = gr.Checkbox(
553
+ label="Open Access Only",
554
+ value=False,
555
+ elem_classes="gr-checkbox"
556
+ )
557
 
558
+ # Year range
559
  with gr.Row():
560
+ with gr.Column():
561
+ year_from = gr.Number(
562
+ label="Year From",
563
+ value=2000,
564
+ minimum=1900,
565
+ maximum=2026,
566
+ elem_classes="gr-input"
567
+ )
568
+
569
+ with gr.Column():
570
+ year_to = gr.Number(
571
+ label="Year To",
572
+ value=2026,
573
+ minimum=1900,
574
+ maximum=2026,
575
+ elem_classes="gr-input"
576
+ )
577
 
578
+ # Search button
579
+ search_btn = gr.Button(
580
+ "πŸ” SEARCH PAPERS",
581
+ variant="primary",
582
+ size="lg",
583
+ elem_classes="gr-button-primary"
584
+ )
585
 
586
+ # Metrics cards (will be updated after search)
587
+ with gr.Row():
588
+ with gr.Column():
589
+ total_papers = gr.Markdown("**0**", elem_classes="metric-value")
590
+ gr.Markdown("Total Papers", elem_classes="metric-label")
591
+
592
+ with gr.Column():
593
+ avg_relevance = gr.Markdown("**0.0**", elem_classes="metric-value")
594
+ gr.Markdown("Avg Relevance", elem_classes="metric-label")
595
+
596
+ with gr.Column():
597
+ year_range = gr.Markdown("**N/A**", elem_classes="metric-value")
598
+ gr.Markdown("Year Range", elem_classes="metric-label")
599
 
600
+ # Results
601
  results = gr.Dataframe(
602
+ label="πŸ“„ Results",
603
+ wrap=True,
604
+ row_count=15,
605
+ elem_classes="gr-dataframe"
606
  )
607
  papers_state = gr.State([])
608
 
609
+ # Export section
 
 
 
 
 
 
 
 
 
 
 
 
 
 
610
  with gr.Row():
611
+ with gr.Column(scale=2):
612
+ filename = gr.Textbox(
613
+ label="Filename",
614
+ value="papers_results",
615
+ placeholder="Enter filename",
616
+ elem_classes="gr-input"
617
+ )
618
+
619
+ with gr.Column(scale=1):
620
+ format_type = gr.Dropdown(
621
+ label="Format",
622
+ choices=["CSV", "Excel"],
623
+ value="CSV",
624
+ elem_classes="gr-input"
625
+ )
626
 
627
+ export_btn = gr.Button(
628
+ "πŸ“₯ DOWNLOAD WITH ABSTRACTS",
629
+ variant="secondary",
630
+ elem_classes="gr-button-secondary"
631
  )
632
+ export_status = gr.Textbox(label="", visible=True)
633
+ download_file = gr.File(label="Download Link", visible=False)
634
 
635
+ # Info box
636
+ gr.HTML("""
637
+ <div class='info-box'>
638
+ <strong>πŸ“Œ Note:</strong> Downloaded files include FULL abstracts.
639
+ Excel format includes a Summary sheet with metadata.
640
+ </div>
641
+ """)
642
+
643
+ # Initialize searcher
644
+ searcher = PaperSearchMobile()
645
+
646
+ # Search function
647
+ def search_wrapper(query, keywords, sort_by, sort_order, max_papers, year_from, year_to, open_access):
648
+ df, papers, years, keyword_list = searcher.search_papers(
649
+ query, keywords, sort_by, sort_order, max_papers, year_from, year_to, open_access
650
  )
651
 
652
+ if papers:
653
+ # Update metrics
654
+ total = len(papers)
655
+ avg_score = sum(p.get('score', 0) for p in papers) / total
656
+ year_min = min(years) if years else 'N/A'
657
+ year_max = max(years) if years else 'N/A'
658
+
659
+ total_md = f"**{total}**"
660
+ avg_md = f"**{avg_score:.1f}**"
661
+ year_md = f"**{year_min} - {year_max}**"
662
+ else:
663
+ total_md = "**0**"
664
+ avg_md = "**0.0**"
665
+ year_md = "**N/A**"
666
+
667
+ return df, papers, total_md, avg_md, year_md
668
+
669
+ search_btn.click(
670
+ search_wrapper,
671
+ inputs=[query, keywords, sort_by, sort_order, max_papers, year_from, year_to, open_access],
672
+ outputs=[results, papers_state, total_papers, avg_relevance, year_range]
673
+ )
674
+
675
+ # Export function
676
+ def export_wrapper(papers, fname, fmt):
677
+ if not papers:
678
+ return None, "No papers to export", gr.File(visible=False)
679
+ file_path, message = download_results(papers, fname, fmt)
680
+ return file_path, message, gr.File(visible=True)
681
+
682
+ export_btn.click(
683
+ export_wrapper,
684
+ inputs=[papers_state, filename, format_type],
685
+ outputs=[download_file, export_status, download_file]
686
+ )
687
 
688
  return demo
689
 
 
692
  # ============================================
693
  if __name__ == "__main__":
694
  demo = create_interface()
 
 
 
 
 
 
 
 
 
 
 
 
695
  demo.launch()