fix format

Author: CodingWithTim

fastchat/constants.py

@@ -25,9 +25,7 @@
 IMAGE_MODERATION_MSG = (
     "$MODERATION$ YOUR IMAGE VIOLATES OUR CONTENT MODERATION GUIDELINES."
 )
-PDF_MODERATION_MSG = (
-    "$MODERATION$ YOUR PDF VIOLATES OUR CONTENT MODERATION GUIDELINES."
-)
+PDF_MODERATION_MSG = "$MODERATION$ YOUR PDF VIOLATES OUR CONTENT MODERATION GUIDELINES."
 MODERATION_MSG = "$MODERATION$ YOUR INPUT VIOLATES OUR CONTENT MODERATION GUIDELINES."
 CONVERSATION_LIMIT_MSG = "YOU HAVE REACHED THE CONVERSATION LENGTH LIMIT. PLEASE CLEAR HISTORY AND START A NEW CONVERSATION."
 INACTIVE_MSG = "THIS SESSION HAS BEEN INACTIVE FOR TOO LONG. PLEASE REFRESH THIS PAGE."

fastchat/serve/gradio_block_arena_vision.py

@@ -265,23 +265,25 @@ def pdf_moderator(images):
     import base64
     from openai import OpenAI
     from io import BytesIO
-    
+
     base64_urls = []
     for image in images:
         buffer = BytesIO()
         image.save(buffer, format="JPEG")
-        
+
         image_bytes = buffer.getvalue()
         image_b64 = base64.b64encode(image_bytes).decode("utf-8")
 
         # convert to openai format
-        base64_urls.append({
-            "type": "image_url",
-            "image_url": {
-                "url": f"data:image/jpeg;base64,{image_b64}",
+        base64_urls.append(
+            {
+                "type": "image_url",
+                "image_url": {
+                    "url": f"data:image/jpeg;base64,{image_b64}",
+                },
             }
-        })
-    
+        )
+
     # OpenAI's maximum number of images is 1 at the moment.
     client = OpenAI()
     moderations = []
@@ -294,7 +296,7 @@ def pdf_moderator(images):
             moderations.append(response[0].results.flagged)
         except Exception as e:
             print(e)
-            
+
     return all(moderations)
 
 
@@ -310,7 +312,7 @@ def detect_language_from_doc(pdf_file_path):
 
     # Convert pdf into image (first page only for efficiency)
     images = convert_from_path(pdf_file_path)
-    
+
     extracted_text = pytesseract.image_to_string(
         images[0], lang=TESSERACT_SUPPORTED_LANGS
     )

fastchat/serve/gradio_block_arena_vision_anony.py

@@ -300,21 +300,13 @@ def add_text(
                 PDFCHAT_SAMPLING_WEIGHTS,
                 PDFCHAT_SAMPLING_BOOST_MODELS,
             )
-            
+
             # Save an unique id for mapping conversation back to the file on google cloud.
             unique_id = hash_pdf(pdfs[0])
-            
+
             states = [
-                State(
-                    model_left, 
-                    is_vision=False, 
-                    pdf_id=unique_id
-                ),
-                State(
-                    model_right, 
-                    is_vision=False, 
-                    pdf_id=unique_id
-                ),
+                State(model_left, is_vision=False, pdf_id=unique_id),
+                State(model_right, is_vision=False, pdf_id=unique_id),
             ]
             upload_pdf_file_to_gcs(
                 pdf_file_path=pdfs[0],

fastchat/serve/gradio_web_server.py

@@ -119,7 +119,9 @@ def __init__(self, model_name, is_vision=False, pdf_id=None):
         self.model_name = model_name
         self.oai_thread_id = None
         self.is_vision = is_vision
-        self.pdf_id = pdf_id # NOTE(Tim): Version 1 PDFChat Architecture, could be revised later.
+        self.pdf_id = (
+            pdf_id  # NOTE(Tim): Version 1 PDFChat Architecture, could be revised later.
+        )
 
         # NOTE(chris): This could be sort of a hack since it assumes the user only uploads one image. If they can upload multiple, we should store a list of image hashes.
         self.has_csam_image = False

fastchat/utils.py

@@ -468,8 +468,8 @@ def get_pdf_file_from_gcs(filename):
 
 def hash_pdf(file_path):
     import hashlib
-    
-    with open(file_path, 'rb') as f:
+
+    with open(file_path, "rb") as f:
         file_content = f.read()
     return hashlib.md5(file_content).hexdigest()