text embeddings for plot

scripts/preprocessing.py

@@ -18,10 +18,6 @@ sym_spell = SymSpell(max_dictionary_edit_distance=2, prefix_length=7)
 stemmer = PorterStemmer()
 lemmatizer = WordNetLemmatizer()
 
-# model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2')
-
-# df = pd.read_excel('C:\\Users\\ishaa\\OneDrive\\Documents\\MSU\\Spring 2026\\Data mining\\Project\\sample_data.xlsx', engine='openpyxl')
-
 def clean_plot(text):
     text = text.lower()
     text = text.translate(str.maketrans('', '', string.punctuation))  # Remove punctuation
@@ -35,11 +31,12 @@ def clean_plot(text):
     return text
 
 def get_genre(row):
+    if pd.isna(row['Genre']):
+        return ""
     movie = row['Title']
-    print(movie)
     text = row['Genre']
-    text = text.split(".")[0]
     text = text.replace(movie, "")
+    text = text.split(".")[0]
     text = text.lower()
     match = re.search(r'is a ((?:\S+\s+){4}\S+)', text)
     if match:
@@ -53,13 +50,14 @@ def get_genre(row):
     return text
 
 def pre_director(text):
-    if not text:
+    if pd.isna(text) or not text:
         return ""
     text = text.lower().strip()
     return text
 
-def clean_cast(text, top_k=5):
+def clean_cast(text):
-    if not text:
+    print(f"Original cast: {text}")
+    if pd.isna(text) or not text:
         return []
     text = text.lower()
     cast_list = [actor.strip() for actor in text.split(",")]

scripts/vec_representation.py

@@ -2,10 +2,12 @@ import pandas as pd
 from sentence_transformers import SentenceTransformer
 from preprocessing import clean_plot, get_genre, pre_director, clean_cast
 from sklearn.metrics.pairwise import cosine_similarity
+import numpy as np
 
-df = pd.read_excel('C:\\Users\\ishaa\\OneDrive\\Documents\\MSU\\Spring 2026\\Data mining\\Project\\updated_data.xlsx', engine='openpyxl')
+df = pd.read_excel('C:\\Users\\ishaa\\OneDrive\\Documents\\MSU\\Spring 2026\\Data mining\\Project\\updated_datav2.xlsx', engine='openpyxl')
+df = df.dropna(subset=['Plot'])
 
-df = df.dropna(subset=['Genre', 'Plot'])
+print(len(df))
 
 df['Processed_Plot'] = df['Plot'].apply(clean_plot)
 
@@ -14,4 +16,18 @@ df['Pre_genre'] = df[['Genre', 'Title']].apply(get_genre, axis=1)
 df['Pre_director'] = df['Director'].apply(pre_director)
 df['Pre_cast'] = df['Cast'].apply(clean_cast)
 
+
+# Load embedding model
+model = SentenceTransformer("all-MiniLM-L6-v2")
+
+# Generate one embedding per movie plot
+plot_embeddings = model.encode(
+    df["Processed_Plot"].tolist(),
+    show_progress_bar=True,
+    convert_to_numpy=True
+)
+
+
+np.save(r"C:\Users\ishaa\OneDrive\Documents\MSU\Spring 2026\Data mining\Project\plot_embeddings.npy",plot_embeddings)
+
 df.to_excel('C:\\Users\\ishaa\\OneDrive\\Documents\\MSU\\Spring 2026\\Data mining\\Project\\preprocessed_data.xlsx', index=False)