Spaces:
Running
Running
Sophie
commited on
Commit
·
c38e4ce
1
Parent(s):
6de8c39
added filtering and citation counts
Browse files- src/streamlit_app.py +242 -39
src/streamlit_app.py
CHANGED
|
@@ -6,6 +6,10 @@ import os
|
|
| 6 |
import boto3
|
| 7 |
import psycopg2
|
| 8 |
from psycopg2.extensions import connection
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
from dotenv import load_dotenv
|
| 10 |
from latex_clean import clean_latex_for_display
|
| 11 |
|
|
@@ -50,6 +54,11 @@ ALLOWED_TYPES = [
|
|
| 50 |
"theorem", "lemma", "proposition", "corollary", "definition", "remark", "assumption"
|
| 51 |
]
|
| 52 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
# Load the Embedding Model
|
| 54 |
@st.cache_resource
|
| 55 |
def load_model():
|
|
@@ -63,7 +72,6 @@ def load_model():
|
|
| 63 |
st.error(f"Error loading the embedding model: {e}")
|
| 64 |
return None
|
| 65 |
|
| 66 |
-
|
| 67 |
# Load Data from RDS
|
| 68 |
@st.cache_data
|
| 69 |
def load_papers_from_rds():
|
|
@@ -129,6 +137,25 @@ def load_papers_from_rds():
|
|
| 129 |
elif isinstance(embedding, np.ndarray):
|
| 130 |
embedding = embedding.astype(np.float32)
|
| 131 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 132 |
all_theorems_data.append({
|
| 133 |
"paper_id": paper_id,
|
| 134 |
"authors": authors,
|
|
@@ -136,11 +163,15 @@ def load_papers_from_rds():
|
|
| 136 |
"paper_url": link,
|
| 137 |
"year": last_updated.year,
|
| 138 |
"primary_category": primary_category,
|
|
|
|
|
|
|
|
|
|
|
|
|
| 139 |
"theorem_name": theorem_name,
|
| 140 |
"theorem_slogan": theorem_slogan,
|
| 141 |
"theorem_body": theorem_body,
|
| 142 |
"global_context": global_context,
|
| 143 |
-
"stored_embedding": embedding
|
| 144 |
})
|
| 145 |
|
| 146 |
return all_theorems_data
|
|
@@ -149,63 +180,181 @@ def load_papers_from_rds():
|
|
| 149 |
st.error(f"Error loading data from RDS: {e}")
|
| 150 |
return []
|
| 151 |
|
| 152 |
-
|
| 153 |
-
|
| 154 |
-
def search_theorems(query, model, theorems_data, embeddings_db):
|
| 155 |
"""
|
| 156 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 157 |
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 158 |
if not query:
|
| 159 |
st.info("Please enter a search query.")
|
| 160 |
return
|
|
|
|
|
|
|
|
|
|
| 161 |
|
| 162 |
query_embedding = model.encode(query, convert_to_tensor=True)
|
| 163 |
cosine_scores = util.cos_sim(query_embedding, embeddings_db)[0]
|
| 164 |
-
top_results_indices = np.argsort(-cosine_scores.cpu())[:10]
|
| 165 |
-
|
| 166 |
-
st.subheader("Top 5 Most Similar Theorems")
|
| 167 |
|
| 168 |
-
|
| 169 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 170 |
return
|
| 171 |
|
| 172 |
-
for i,
|
| 173 |
-
|
| 174 |
-
|
| 175 |
-
theorem_info = theorems_data[idx]
|
| 176 |
-
|
| 177 |
-
expander_title = f"**Result {i+1} | Similarity: {similarity:.4f}**"
|
| 178 |
-
if theorem_info.get("theorem_name"):
|
| 179 |
-
expander_title += f" | {theorem_info['theorem_name']}"
|
| 180 |
-
|
| 181 |
with st.expander(expander_title):
|
| 182 |
-
st.markdown(f"**Paper:** {
|
| 183 |
-
st.markdown(f"**Authors:** {', '.join(
|
| 184 |
-
st.markdown(f"**Source:**
|
|
|
|
|
|
|
| 185 |
st.markdown(
|
| 186 |
-
f"**Math Tag:** `{
|
|
|
|
|
|
|
|
|
|
| 187 |
st.markdown("---")
|
| 188 |
|
| 189 |
-
if
|
| 190 |
-
st.markdown(f"**Slogan:** {
|
| 191 |
-
st.write("")
|
| 192 |
|
| 193 |
-
if
|
| 194 |
-
cleaned_ctx = clean_latex_for_display(
|
| 195 |
-
|
| 196 |
-
st.markdown(blockquote_context)
|
| 197 |
-
st.write("")
|
| 198 |
|
| 199 |
-
cleaned_content = clean_latex_for_display(
|
| 200 |
-
st.markdown(
|
| 201 |
st.markdown(cleaned_content)
|
| 202 |
|
| 203 |
# --- Main App Interface ---
|
| 204 |
st.set_page_config(page_title="Theorem Search Demo", layout="wide")
|
| 205 |
st.title("📚 Semantic Theorem Search")
|
| 206 |
st.write("This demo uses a specialized mathematical language model to find theorems semantically similar to your query.")
|
| 207 |
-
st.markdown("*Note: Linking to a specific page within an arXiv PDF is not directly possible.*",
|
| 208 |
-
help="arXiv links redirect to the paper's abstract, not a specific page in the PDF.")
|
| 209 |
|
| 210 |
model = load_model()
|
| 211 |
theorems_data = load_papers_from_rds()
|
|
@@ -214,11 +363,65 @@ if model and theorems_data:
|
|
| 214 |
with st.spinner("Preparing embeddings from database..."):
|
| 215 |
corpus_embeddings = np.array([item['stored_embedding'] for item in theorems_data])
|
| 216 |
|
| 217 |
-
st.success(f"Successfully loaded {len(theorems_data)} theorems from arXiv. Ready to search!")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 218 |
|
| 219 |
user_query = st.text_input("Enter your query:", "")
|
| 220 |
-
|
| 221 |
if st.button("Search") or user_query:
|
| 222 |
-
|
| 223 |
else:
|
| 224 |
st.error("Could not load the model or data from RDS. Please check your RDS database connection and credentials.")
|
|
|
|
| 6 |
import boto3
|
| 7 |
import psycopg2
|
| 8 |
from psycopg2.extensions import connection
|
| 9 |
+
import torch
|
| 10 |
+
import re
|
| 11 |
+
import requests
|
| 12 |
+
from concurrent.futures import ThreadPoolExecutor, as_completed
|
| 13 |
from dotenv import load_dotenv
|
| 14 |
from latex_clean import clean_latex_for_display
|
| 15 |
|
|
|
|
| 54 |
"theorem", "lemma", "proposition", "corollary", "definition", "remark", "assumption"
|
| 55 |
]
|
| 56 |
|
| 57 |
+
ARXIV_ID_RE = re.compile(
|
| 58 |
+
r'arxiv\.org/(?:abs|pdf)/((?:\d{4}\.\d{4,5}|[a-z\-]+/\d{7}))(?:v\d+)?',
|
| 59 |
+
re.IGNORECASE
|
| 60 |
+
)
|
| 61 |
+
|
| 62 |
# Load the Embedding Model
|
| 63 |
@st.cache_resource
|
| 64 |
def load_model():
|
|
|
|
| 72 |
st.error(f"Error loading the embedding model: {e}")
|
| 73 |
return None
|
| 74 |
|
|
|
|
| 75 |
# Load Data from RDS
|
| 76 |
@st.cache_data
|
| 77 |
def load_papers_from_rds():
|
|
|
|
| 137 |
elif isinstance(embedding, np.ndarray):
|
| 138 |
embedding = embedding.astype(np.float32)
|
| 139 |
|
| 140 |
+
# Determine source from url
|
| 141 |
+
link_str = link or ""
|
| 142 |
+
if link_str.startswith("http://arxiv.org") or link_str.startswith("https://arxiv.org"):
|
| 143 |
+
source = "arXiv"
|
| 144 |
+
else:
|
| 145 |
+
source = "Stacks Project"
|
| 146 |
+
|
| 147 |
+
# Determine type from name
|
| 148 |
+
def infer_type(name: str) -> str:
|
| 149 |
+
if not name:
|
| 150 |
+
return "theorem"
|
| 151 |
+
lower = name.lower()
|
| 152 |
+
for t in ["theorem", "lemma", "proposition", "corollary", "definition", "remark", "assumption"]:
|
| 153 |
+
if t in lower:
|
| 154 |
+
return t
|
| 155 |
+
return "theorem"
|
| 156 |
+
|
| 157 |
+
inferred_type = infer_type(theorem_name or "")
|
| 158 |
+
|
| 159 |
all_theorems_data.append({
|
| 160 |
"paper_id": paper_id,
|
| 161 |
"authors": authors,
|
|
|
|
| 163 |
"paper_url": link,
|
| 164 |
"year": last_updated.year,
|
| 165 |
"primary_category": primary_category,
|
| 166 |
+
"source": source,
|
| 167 |
+
"type": inferred_type,
|
| 168 |
+
"journal_published": bool(journal_ref),
|
| 169 |
+
"citations": None,
|
| 170 |
"theorem_name": theorem_name,
|
| 171 |
"theorem_slogan": theorem_slogan,
|
| 172 |
"theorem_body": theorem_body,
|
| 173 |
"global_context": global_context,
|
| 174 |
+
"stored_embedding": embedding,
|
| 175 |
})
|
| 176 |
|
| 177 |
return all_theorems_data
|
|
|
|
| 180 |
st.error(f"Error loading data from RDS: {e}")
|
| 181 |
return []
|
| 182 |
|
| 183 |
+
@st.cache_data(ttl=60*60*24) # cache for 24 hours
|
| 184 |
+
def fetch_citations(paper_url: str, title: str) -> int | None:
|
|
|
|
| 185 |
"""
|
| 186 |
+
Returns citation count if found, else None.
|
| 187 |
+
Tries the following sources in order:
|
| 188 |
+
1) OpenAlex by arXiv id
|
| 189 |
+
2) Semantic Scholar by arXiv id
|
| 190 |
+
3) Semantic Scholar by title
|
| 191 |
"""
|
| 192 |
+
arx_id = None
|
| 193 |
+
if paper_url:
|
| 194 |
+
m = ARXIV_ID_RE.search(paper_url)
|
| 195 |
+
if m:
|
| 196 |
+
arx_id = m.group(1)
|
| 197 |
+
# OpenAlex by arXiv id
|
| 198 |
+
if arx_id:
|
| 199 |
+
try:
|
| 200 |
+
r = requests.get(f"https://api.openalex.org/works/arXiv:{arx_id}", timeout=10)
|
| 201 |
+
if r.ok:
|
| 202 |
+
data = r.json()
|
| 203 |
+
c = data.get("cited_by_count")
|
| 204 |
+
if isinstance(c, int):
|
| 205 |
+
return c
|
| 206 |
+
except Exception:
|
| 207 |
+
pass
|
| 208 |
+
# Semantic Scholar by arXiv id
|
| 209 |
+
if arx_id:
|
| 210 |
+
try:
|
| 211 |
+
r = requests.get(
|
| 212 |
+
f"https://api.semanticscholar.org/graph/v1/paper/arXiv:{arx_id}",
|
| 213 |
+
params={"fields": "citationCount"},
|
| 214 |
+
timeout=10
|
| 215 |
+
)
|
| 216 |
+
if r.ok:
|
| 217 |
+
j = r.json()
|
| 218 |
+
c = j.get("citationCount")
|
| 219 |
+
if isinstance(c, int):
|
| 220 |
+
return c
|
| 221 |
+
except Exception:
|
| 222 |
+
pass
|
| 223 |
+
# Fallback: Semantic Scholar by title
|
| 224 |
+
if title:
|
| 225 |
+
try:
|
| 226 |
+
r = requests.get(
|
| 227 |
+
"https://api.semanticscholar.org/graph/v1/paper/search",
|
| 228 |
+
params={"query": title, "limit": 1, "fields": "title,citationCount"},
|
| 229 |
+
timeout=10
|
| 230 |
+
)
|
| 231 |
+
if r.ok:
|
| 232 |
+
j = r.json()
|
| 233 |
+
if j.get("data"):
|
| 234 |
+
c = j["data"][0].get("citationCount")
|
| 235 |
+
if isinstance(c, int):
|
| 236 |
+
return c
|
| 237 |
+
except Exception:
|
| 238 |
+
pass
|
| 239 |
+
|
| 240 |
+
return None
|
| 241 |
+
|
| 242 |
+
def add_citations(candidates: list[dict], max_workers: int = 6) -> None:
|
| 243 |
+
# Select targets with missing citations
|
| 244 |
+
targets = [
|
| 245 |
+
it for it in candidates
|
| 246 |
+
if it.get("source") == "arXiv" and (it.get("citations") in (None, 0))
|
| 247 |
+
]
|
| 248 |
+
if not targets:
|
| 249 |
+
return
|
| 250 |
+
|
| 251 |
+
with ThreadPoolExecutor(max_workers=max_workers) as exe:
|
| 252 |
+
fut2item = {
|
| 253 |
+
exe.submit(fetch_citations, it.get("paper_url"), it.get("paper_title")): it
|
| 254 |
+
for it in targets
|
| 255 |
+
}
|
| 256 |
+
for fut in as_completed(fut2item):
|
| 257 |
+
it = fut2item[fut]
|
| 258 |
+
try:
|
| 259 |
+
c = fut.result()
|
| 260 |
+
if c is not None:
|
| 261 |
+
it["citations"] = c
|
| 262 |
+
except Exception:
|
| 263 |
+
pass
|
| 264 |
+
|
| 265 |
+
# --- Search and Display ---
|
| 266 |
+
def search_and_display_with_filters(query, model, theorems_data, embeddings_db, filters):
|
| 267 |
if not query:
|
| 268 |
st.info("Please enter a search query.")
|
| 269 |
return
|
| 270 |
+
if not filters['sources']:
|
| 271 |
+
st.warning("Please select at least one source.")
|
| 272 |
+
return
|
| 273 |
|
| 274 |
query_embedding = model.encode(query, convert_to_tensor=True)
|
| 275 |
cosine_scores = util.cos_sim(query_embedding, embeddings_db)[0]
|
|
|
|
|
|
|
|
|
|
| 276 |
|
| 277 |
+
# Get a larger pool to filter from
|
| 278 |
+
top_k_pool = min(200, len(theorems_data))
|
| 279 |
+
top_indices = torch.topk(cosine_scores, k=top_k_pool, sorted=True).indices
|
| 280 |
+
pool_items = [theorems_data[int(i.item())] for i in top_indices]
|
| 281 |
+
add_citations(pool_items)
|
| 282 |
+
|
| 283 |
+
results = []
|
| 284 |
+
low, high = filters['citation_range']
|
| 285 |
+
|
| 286 |
+
# Filter results
|
| 287 |
+
for item in pool_items:
|
| 288 |
+
type_match = (not filters['types']) or (item.get('type','').lower() in filters['types'])
|
| 289 |
+
tag_match = (not filters['tags']) or (item.get('primary_category') in filters['tags'])
|
| 290 |
+
author_match = (not filters['authors']) or any(a in (item.get('authors') or []) for a in filters['authors'])
|
| 291 |
+
source_match = item.get('source') in filters['sources']
|
| 292 |
+
|
| 293 |
+
# Citations & year & journal only meaningful for arXiv
|
| 294 |
+
cit = item.get('citations')
|
| 295 |
+
if cit is None:
|
| 296 |
+
if not filters['include_unknown_citations']:
|
| 297 |
+
continue
|
| 298 |
+
citation_match = True
|
| 299 |
+
else:
|
| 300 |
+
citation_match = (low <= int(cit) <= high)
|
| 301 |
+
|
| 302 |
+
year_match = True
|
| 303 |
+
if filters['year_range'] and item.get('source') == 'arXiv':
|
| 304 |
+
y = item.get('year') or 0
|
| 305 |
+
yr0, yr1 = filters['year_range']
|
| 306 |
+
year_match = (yr0 <= y <= yr1)
|
| 307 |
+
|
| 308 |
+
journal_match = True
|
| 309 |
+
if item.get('source') == 'arXiv':
|
| 310 |
+
status = filters['journal_status']
|
| 311 |
+
jp = bool(item.get('journal_published'))
|
| 312 |
+
if status == "Journal Article":
|
| 313 |
+
journal_match = jp
|
| 314 |
+
elif status == "Preprint Only":
|
| 315 |
+
journal_match = not jp
|
| 316 |
+
|
| 317 |
+
if all([type_match, tag_match, author_match, source_match, citation_match, year_match, journal_match]):
|
| 318 |
+
results.append({"info": item, "similarity": float(cosine_scores[theorems_data.index(item)].item())})
|
| 319 |
+
if len(results) >= filters['top_k']:
|
| 320 |
+
break
|
| 321 |
+
|
| 322 |
+
st.subheader(f"Found {len(results)} Matching Results")
|
| 323 |
+
if not results:
|
| 324 |
+
st.warning("No results found for the current filters.")
|
| 325 |
return
|
| 326 |
|
| 327 |
+
for i, r in enumerate(results):
|
| 328 |
+
info = r["info"]
|
| 329 |
+
expander_title = f"**Result {i+1} | Similarity: {r['similarity']:.4f} | Type: {info.get('type','').title()}**"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 330 |
with st.expander(expander_title):
|
| 331 |
+
st.markdown(f"**Paper:** *{info.get('paper_title','Unknown')}*")
|
| 332 |
+
st.markdown(f"**Authors:** {', '.join(info.get('authors') or []) or 'N/A'}")
|
| 333 |
+
st.markdown(f"**Source:** {info.get('source')} ([Link]({info.get('paper_url')}))")
|
| 334 |
+
cit = info.get("citations")
|
| 335 |
+
cit_str = "Unknown" if cit is None else str(cit)
|
| 336 |
st.markdown(
|
| 337 |
+
f"**Math Tag:** `{info.get('primary_category')}` | "
|
| 338 |
+
f"**Citations:** {cit_str} | "
|
| 339 |
+
f"**Year:** {info.get('year', 'N/A')}"
|
| 340 |
+
)
|
| 341 |
st.markdown("---")
|
| 342 |
|
| 343 |
+
if info.get("theorem_slogan"):
|
| 344 |
+
st.markdown(f"**Slogan:** {info['theorem_slogan']}\n")
|
|
|
|
| 345 |
|
| 346 |
+
if info.get("global_context"):
|
| 347 |
+
cleaned_ctx = clean_latex_for_display(info["global_context"])
|
| 348 |
+
st.markdown("> " + cleaned_ctx.replace("\n", "\n> ") )
|
|
|
|
|
|
|
| 349 |
|
| 350 |
+
cleaned_content = clean_latex_for_display(info['theorem_body'])
|
| 351 |
+
st.markdown("**Theorem Body:**")
|
| 352 |
st.markdown(cleaned_content)
|
| 353 |
|
| 354 |
# --- Main App Interface ---
|
| 355 |
st.set_page_config(page_title="Theorem Search Demo", layout="wide")
|
| 356 |
st.title("📚 Semantic Theorem Search")
|
| 357 |
st.write("This demo uses a specialized mathematical language model to find theorems semantically similar to your query.")
|
|
|
|
|
|
|
| 358 |
|
| 359 |
model = load_model()
|
| 360 |
theorems_data = load_papers_from_rds()
|
|
|
|
| 363 |
with st.spinner("Preparing embeddings from database..."):
|
| 364 |
corpus_embeddings = np.array([item['stored_embedding'] for item in theorems_data])
|
| 365 |
|
| 366 |
+
st.success(f"Successfully loaded {len(theorems_data)} theorems from arXiv and the Stacks Project. Ready to search!")
|
| 367 |
+
|
| 368 |
+
# --- Sidebar filters ---
|
| 369 |
+
with st.sidebar:
|
| 370 |
+
st.header("Search Filters")
|
| 371 |
+
|
| 372 |
+
all_sources = ['arXiv', 'Stacks Project']
|
| 373 |
+
selected_sources = st.multiselect(
|
| 374 |
+
"Filter by Source(s):",
|
| 375 |
+
all_sources,
|
| 376 |
+
default=all_sources[:1] if all_sources else [],
|
| 377 |
+
help="Select one or more sources to reveal more filters."
|
| 378 |
+
)
|
| 379 |
+
|
| 380 |
+
selected_authors, selected_types, selected_tags = [], [], []
|
| 381 |
+
year_range, journal_status = None, "All"
|
| 382 |
+
citation_range = (0, 1000)
|
| 383 |
+
top_k_results = 5
|
| 384 |
+
|
| 385 |
+
if selected_sources:
|
| 386 |
+
st.write("---")
|
| 387 |
+
selected_types = st.multiselect("Filter by Type:", ALLOWED_TYPES)
|
| 388 |
+
all_authors = sorted(list(set(a for it in theorems_data for a in (it.get('authors') or []))))
|
| 389 |
+
selected_authors = st.multiselect("Filter by Author(s):", all_authors)
|
| 390 |
+
|
| 391 |
+
# Tags come from union of categories per selected source
|
| 392 |
+
from collections import defaultdict
|
| 393 |
+
tags_per_source = defaultdict(set)
|
| 394 |
+
for it in theorems_data:
|
| 395 |
+
tags_per_source[it['source']].add(it.get('primary_category'))
|
| 396 |
+
union_tags = sorted({t for s in selected_sources for t in tags_per_source.get(s, set()) if t})
|
| 397 |
+
selected_tags = st.multiselect("Filter by Math Tag/Category:", union_tags)
|
| 398 |
+
|
| 399 |
+
if 'arXiv' in selected_sources:
|
| 400 |
+
year_range = st.slider("Filter by Year (for arXiv):", 1991, 2025, (1991, 2025))
|
| 401 |
+
journal_status = st.radio("Publication Status (for arXiv):", ["All", "Journal Article", "Preprint Only"], horizontal=True)
|
| 402 |
+
|
| 403 |
+
citation_range = st.slider("Filter by Citations:", 0, 1000, (0, 1000))
|
| 404 |
+
include_unknown_citations = st.checkbox(
|
| 405 |
+
"Include entries with unknown citation counts",
|
| 406 |
+
value=True,
|
| 407 |
+
help="If unchecked, results with unknown citation counts are excluded."
|
| 408 |
+
)
|
| 409 |
+
top_k_results = st.slider("Number of results to display:", 1, 20, 5)
|
| 410 |
+
|
| 411 |
+
filters = {
|
| 412 |
+
"authors": selected_authors,
|
| 413 |
+
"types": [t.lower() for t in selected_types],
|
| 414 |
+
"tags": selected_tags,
|
| 415 |
+
"sources": selected_sources,
|
| 416 |
+
"year_range": year_range,
|
| 417 |
+
"journal_status": journal_status,
|
| 418 |
+
"citation_range": citation_range,
|
| 419 |
+
"include_unknown_citations": include_unknown_citations,
|
| 420 |
+
"top_k": top_k_results
|
| 421 |
+
}
|
| 422 |
|
| 423 |
user_query = st.text_input("Enter your query:", "")
|
|
|
|
| 424 |
if st.button("Search") or user_query:
|
| 425 |
+
search_and_display_with_filters(user_query, model, theorems_data, corpus_embeddings, filters)
|
| 426 |
else:
|
| 427 |
st.error("Could not load the model or data from RDS. Please check your RDS database connection and credentials.")
|