Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -8,24 +8,22 @@ from transformers import CLIPProcessor, CLIPModel
|
|
8 |
from PIL import Image
|
9 |
import chromadb
|
10 |
import numpy as np
|
|
|
11 |
|
12 |
app = FastAPI()
|
13 |
|
14 |
-
# Initialize ChromaDB
|
15 |
client = chromadb.PersistentClient(path="/data/chroma_db")
|
16 |
-
client.
|
17 |
-
collection = client.get_or_create_collection(name="knowledge_base", metadata={"dim": 512})
|
18 |
-
|
19 |
-
#collection = client.get_or_create_collection(name="knowledge_base", metadata={"hnsw:space": "cosine"}, embedding_function=None)
|
20 |
|
21 |
# File Paths
|
22 |
pdf_file = "Sutures and Suturing techniques.pdf"
|
23 |
pptx_file = "impalnt 1.pptx"
|
24 |
|
25 |
# Initialize Embedding Models
|
26 |
-
text_model = SentenceTransformer('
|
27 |
-
|
28 |
-
|
29 |
|
30 |
# Image Storage Folder
|
31 |
IMAGE_FOLDER = "/data/extracted_images"
|
@@ -88,17 +86,23 @@ def extract_images_from_pptx(pptx_path):
|
|
88 |
print(f"Error extracting images from PPTX: {e}")
|
89 |
return []
|
90 |
|
91 |
-
# Convert Text to Embeddings
|
92 |
def get_text_embedding(text):
|
93 |
return text_model.encode(text).tolist()
|
94 |
|
95 |
-
# Extract Image Embeddings
|
96 |
def get_image_embedding(image_path):
|
97 |
try:
|
98 |
image = Image.open(image_path)
|
99 |
-
inputs =
|
100 |
with torch.no_grad():
|
101 |
-
image_embedding =
|
|
|
|
|
|
|
|
|
|
|
|
|
102 |
return image_embedding.tolist()
|
103 |
except Exception as e:
|
104 |
print(f"Error generating image embedding: {e}")
|
@@ -109,12 +113,21 @@ def store_data(texts, image_paths):
|
|
109 |
for i, text in enumerate(texts):
|
110 |
if text:
|
111 |
text_embedding = get_text_embedding(text)
|
112 |
-
|
|
|
|
|
|
|
113 |
|
114 |
-
|
115 |
-
|
116 |
-
|
117 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
118 |
|
119 |
print("Data stored successfully!")
|
120 |
|
@@ -133,6 +146,8 @@ def process_and_store(pdf_path=None, pptx_path=None):
|
|
133 |
images.extend(extract_images_from_pptx(pptx_path))
|
134 |
store_data(texts, images)
|
135 |
|
|
|
|
|
136 |
# FastAPI Endpoints
|
137 |
@app.get("/")
|
138 |
def greet_json():
|
|
|
8 |
from PIL import Image
|
9 |
import chromadb
|
10 |
import numpy as np
|
11 |
+
from sklearn.decomposition import PCA
|
12 |
|
13 |
app = FastAPI()
|
14 |
|
15 |
+
# Initialize ChromaDB
|
16 |
client = chromadb.PersistentClient(path="/data/chroma_db")
|
17 |
+
collection = client.get_or_create_collection(name="knowledge_base")
|
|
|
|
|
|
|
18 |
|
19 |
# File Paths
|
20 |
pdf_file = "Sutures and Suturing techniques.pdf"
|
21 |
pptx_file = "impalnt 1.pptx"
|
22 |
|
23 |
# Initialize Embedding Models
|
24 |
+
text_model = SentenceTransformer('all-MiniLM-L6-v2')
|
25 |
+
model = CLIPModel.from_pretrained("openai/clip-vit-base-patch32")
|
26 |
+
processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
|
27 |
|
28 |
# Image Storage Folder
|
29 |
IMAGE_FOLDER = "/data/extracted_images"
|
|
|
86 |
print(f"Error extracting images from PPTX: {e}")
|
87 |
return []
|
88 |
|
89 |
+
# Convert Text to Embeddings
|
90 |
def get_text_embedding(text):
|
91 |
return text_model.encode(text).tolist()
|
92 |
|
93 |
+
# Extract Image Embeddings and Reduce to 384 Dimensions
|
94 |
def get_image_embedding(image_path):
|
95 |
try:
|
96 |
image = Image.open(image_path)
|
97 |
+
inputs = processor(images=image, return_tensors="pt")
|
98 |
with torch.no_grad():
|
99 |
+
image_embedding = model.get_image_features(**inputs).numpy().flatten()
|
100 |
+
|
101 |
+
# Ensure embedding is 384-dimensional
|
102 |
+
if len(image_embedding) != 384:
|
103 |
+
pca = PCA(n_components=384)
|
104 |
+
image_embedding = pca.fit_transform(image_embedding.reshape(1, -1)).flatten()
|
105 |
+
|
106 |
return image_embedding.tolist()
|
107 |
except Exception as e:
|
108 |
print(f"Error generating image embedding: {e}")
|
|
|
113 |
for i, text in enumerate(texts):
|
114 |
if text:
|
115 |
text_embedding = get_text_embedding(text)
|
116 |
+
if len(text_embedding) == 384:
|
117 |
+
collection.add(ids=[f"text_{i}"], embeddings=[text_embedding], documents=[text])
|
118 |
+
|
119 |
+
all_embeddings = [get_image_embedding(img_path) for img_path in image_paths if get_image_embedding(img_path) is not None]
|
120 |
|
121 |
+
if all_embeddings:
|
122 |
+
all_embeddings = np.array(all_embeddings)
|
123 |
+
|
124 |
+
# Apply PCA only if necessary
|
125 |
+
if all_embeddings.shape[1] != 384:
|
126 |
+
pca = PCA(n_components=384)
|
127 |
+
all_embeddings = pca.fit_transform(all_embeddings)
|
128 |
+
|
129 |
+
for j, img_path in enumerate(image_paths):
|
130 |
+
collection.add(ids=[f"image_{j}"], embeddings=[all_embeddings[j].tolist()], documents=[img_path])
|
131 |
|
132 |
print("Data stored successfully!")
|
133 |
|
|
|
146 |
images.extend(extract_images_from_pptx(pptx_path))
|
147 |
store_data(texts, images)
|
148 |
|
149 |
+
|
150 |
+
|
151 |
# FastAPI Endpoints
|
152 |
@app.get("/")
|
153 |
def greet_json():
|