Embedded Files
This project created a Python-based Instagram scraper to gather user interactions, reel metadata, and comments. Using machine learning methods, the data was analyzed for sentiment analysis and multilingual hate speech identification. NLP, web scraping, and sophisticated visualization techniques are all included into the project.
Important code snippets:
Important code snippets:
Scraping Instagram Comments with caption:
def get_ig_comments(post_url, username):
L = instaloader.Instaloader() L.load_session_from_file(username, f"/Users/(username)/.config/instaloader/session-{username}") shortcode = post_url.split("/")[-2] post = instaloader.Post.from_shortcode(L.context, shortcode) caption = post.caption if post.caption else "No caption" comments_data = []
for count, comment in enumerate(post.get_comments()): comments_data.append((comment.owner.username, comment.text,caption))
if (count + 1) % 100 == 0: time.sleep(60) return comments_data
L = instaloader.Instaloader() L.load_session_from_file(username, f"/Users/(username)/.config/instaloader/session-{username}") shortcode = post_url.split("/")[-2] post = instaloader.Post.from_shortcode(L.context, shortcode) caption = post.caption if post.caption else "No caption" comments_data = []
for count, comment in enumerate(post.get_comments()): comments_data.append((comment.owner.username, comment.text,caption))
if (count + 1) % 100 == 0: time.sleep(60) return comments_data
Results of this script:
Fetches Post Captions, Scrapes Comments by automatically pauses after every 100 comments to comply with Instagram’s API rate limits and saves the extracted data in a structured CSV file for easy analysis.
Extracting Text from Images (OCR):
def process_image_for_description(image_path): """Extract and clean text from image using Tesseract OCR.""" image = cv2.imread(image_path) if image is None: return "Error: Unable to load image."
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY) _, thresh = cv2.threshold(gray, 150, 255, cv2.THRESH_BINARY) inverted = cv2.bitwise_not(thresh)
custom_config = r'--oem 3 --psm 6' raw_text = pytesseract.image_to_string(inverted, config=custom_config)
return raw_text if raw_text.strip() else "No readable text found in the image."
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY) _, thresh = cv2.threshold(gray, 150, 255, cv2.THRESH_BINARY) inverted = cv2.bitwise_not(thresh)
custom_config = r'--oem 3 --psm 6' raw_text = pytesseract.image_to_string(inverted, config=custom_config)
return raw_text if raw_text.strip() else "No readable text found in the image."
Conversion of Image from text using BLIP:
def process_image_for_blip_description(image_path): """Generate a caption for the image using the BLIP model.""" try: image = Image.open(image_path) except Exception as e: return f"Error: Unable to load image. {str(e)}"
inputs = processor(images=image, return_tensors="pt") with torch.no_grad(): out = caption_model.generate(**inputs) description = processor.decode(out[0], skip_special_tokens=True) return description
inputs = processor(images=image, return_tensors="pt") with torch.no_grad(): out = caption_model.generate(**inputs) description = processor.decode(out[0], skip_special_tokens=True) return description
Downloading reels , extracting audio and converting to text:
Addendum with report.pdfIf interested to view the report , please request access : )
Page updated
Report abuse