|
import pandas as pd
|
|
import torch
|
|
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
|
from pathlib import Path
|
|
import nltk
|
|
from tqdm import tqdm
|
|
import sys, os.path
|
|
|
|
nltk.download('punkt')
|
|
|
|
BASE_DIR = Path(__file__).resolve().parent.parent.parent
|
|
sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), '../../')))
|
|
|
|
from glimpse.glimpse.data_loading.Glimpse_tokenizer import glimpse_tokenizer
|
|
|
|
|
|
|
|
MODEL_DIR = BASE_DIR / "scibert" / "scibert_topic" / "final_model"
|
|
DATA_DIR = BASE_DIR / "glimpse" / "data" / "processed"
|
|
OUTPUT_DIR = BASE_DIR / "data" / "topic_scored"
|
|
OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
|
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(MODEL_DIR)
|
|
model = AutoModelForSequenceClassification.from_pretrained(MODEL_DIR)
|
|
model.eval()
|
|
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
|
model.to(device)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
id2label = {
|
|
|
|
|
|
|
|
|
|
|
|
|
|
0: "Substance",
|
|
1: "Clarity",
|
|
2: "Soundness/Correctness",
|
|
3: "Originality",
|
|
4: "Motivation/Impact",
|
|
5: "Meaningful Comparison",
|
|
6: "Replicability",
|
|
7: "NONE"
|
|
}
|
|
|
|
def predict_topic(sentences):
|
|
inputs = tokenizer(sentences, return_tensors="pt", padding=True, truncation=True, max_length=512).to(device)
|
|
with torch.no_grad():
|
|
outputs = model(**inputs)
|
|
predictions = torch.argmax(outputs.logits, dim=1).cpu().tolist()
|
|
|
|
predictions = [id2label[pred] for pred in predictions]
|
|
return predictions
|
|
|
|
|
|
def find_topic(start_year=2017, end_year=2021):
|
|
for year in range(start_year, end_year + 1):
|
|
print(f"Processing {year}...")
|
|
input_path = DATA_DIR / f"all_reviews_{year}.csv"
|
|
output_path = OUTPUT_DIR / f"topic_scored_reviews_{year}.csv"
|
|
|
|
df = pd.read_csv(input_path)
|
|
|
|
all_rows = []
|
|
for _, row in tqdm(df.iterrows(), total=len(df)):
|
|
review_id = row["id"]
|
|
text = row["text"]
|
|
sentences = glimpse_tokenizer(text)
|
|
if not sentences:
|
|
continue
|
|
labels = predict_topic(sentences)
|
|
for sentence, topic in zip(sentences, labels):
|
|
all_rows.append({"id": review_id, "sentence": sentence, "topic": topic})
|
|
|
|
output_df = pd.DataFrame(all_rows)
|
|
output_df.to_csv(output_path, index=False)
|
|
print(f"Saved topic-scored data to {output_path}")
|
|
|
|
|
|
if __name__ == "__main__":
|
|
find_topic() |