Spaces:
Sleeping
Sleeping
File size: 1,157 Bytes
00ff675 | 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 | from sklearn.metrics.pairwise import cosine_similarity
from semantic_engine.embed import embed
SUBJECT_ANCHORS = {
"DSA": [
"data structures algorithms coding problems complexity"
],
"OS": [
"operating system deadlock process memory scheduling"
],
"DBMS": [
"database sql normalization indexing transactions"
],
"CLOUD / DEVOPS": [
"cloud aws gcp azure deployment scalability"
],
"FRONTEND": [
"react frontend state management ui"
],
"SYSTEM DESIGN": [
"system design architecture scalability"
],
"HR": [
"hr behavioral teamwork relocation goals"
],
}
# Precompute anchor embeddings
ANCHOR_EMBEDS = {
k: embed(v) for k, v in SUBJECT_ANCHORS.items()
}
def classify_unit(text: str):
unit_emb = embed([text])[0]
best_subject = None
best_score = 0.0
for subject, emb in ANCHOR_EMBEDS.items():
score = cosine_similarity(
[unit_emb], emb
).max()
if score > best_score:
best_score = score
best_subject = subject
return best_subject, round(float(best_score), 2)
|