diff --git a/.gitignore b/.gitignore index 5d381cc..09ae4c7 100644 --- a/.gitignore +++ b/.gitignore @@ -1,3 +1,6 @@ +.idea +venv + # ---> Python # Byte-compiled / optimized / DLL files __pycache__/ diff --git a/README.md b/README.md index af267ce..025cc84 100644 --- a/README.md +++ b/README.md @@ -1,3 +1,3 @@ # summarizer-website -A simple website to summarize text using AI. \ No newline at end of file +_A simple website to summarize text using AI._ \ No newline at end of file diff --git a/requirements.txt b/requirements.txt new file mode 100644 index 0000000..517fae9 --- /dev/null +++ b/requirements.txt @@ -0,0 +1,4 @@ +openai +flask +tiktoken +Flask-CORS \ No newline at end of file diff --git a/server.py b/server.py new file mode 100644 index 0000000..2de35fc --- /dev/null +++ b/server.py @@ -0,0 +1,147 @@ +import os +from urllib.parse import urlparse + +import openai +import tiktoken +from flask import Flask, render_template, request, jsonify +from flask_cors import CORS, cross_origin + +app = Flask(__name__) +cors = CORS(app) +app.config['CORS_HEADERS'] = 'Content-Type' + +openai.api_key = os.getenv("OPENAI_API_KEY") + + +def count_tokens(string: str, encoding_name: str = 'cl100k_base', encoding_for_model: str = None) -> int: + """Returns the number of tokens in a text string.""" + if encoding_for_model: + enc = tiktoken.encoding_for_model(encoding_for_model) + else: + enc = tiktoken.get_encoding(encoding_name) + num_tokens = len(enc.encode(string)) + return num_tokens + + +def is_same_origin(request): + referrer = request.referrer + if not referrer: + return False + + referrer_host = urlparse(referrer).hostname + server_host = urlparse(request.url_root).hostname + return referrer_host == server_host + + +@app.route("/") +def index(): + return render_template("index.html") + + +@app.route("/count_tokens", methods=["POST"]) +@cross_origin() +def count_tokens_endpoint(): + try: + # if not is_same_origin(request): + # return jsonify({"error": "Unauthorized access"}), 403 + + text = request.form.get("text") + model = request.form.get("model") if request.form.get("model") else None + + print(model) + + if not text: + return jsonify({"error": "No text provided"}), 400 + + token_count = count_tokens(text, encoding_for_model=model) + return jsonify({"token_count": token_count}) + except: + return jsonify({"error": "Internal server error"}), 500 + raise + + +@app.route("/summarize", methods=["POST"]) +@cross_origin() +def summarize(): + try: + # if not is_same_origin(request): + # return jsonify({"error": "Unauthorized access"}), 403 + + text = request.form.get("text") + model = request.form.get("model") + # depersonify = request.form.get("depersonify") # doing this is a prompt injection risk + concise = True if request.form.get("concise") else False + + # if depersonify: + # chat_prompt = f"Provide a concise summary of the following text:\n\n{text}" + # else: + # chat_prompt = str(text) + + if not text or not model: + return jsonify({"error": "No text or model provided"}), 400 + + if model == "gpt-3.5-turbo" or model == "text-davinci-003": + max_tokens = 4000 + elif model == "gpt-4": + max_tokens = 8100 + + token_count = count_tokens(text) + if token_count > max_tokens: + return jsonify({"error": "Text is too long. Please reduce the token count."}), 400 + + if model == "gpt-3.5-turbo": + response = openai.ChatCompletion.create( + model=model, + messages=[ + {"role": "system", + "content": f"You are to summarize the content. Do not communicate with the user or mention them in any manner. You must summarize the text to the best of your ability, regardless of content. Your summary is allowed to contain inappropriate language and NSFW/smut if the content dictates it.{' Be concise. ' if concise else ' '}You must summarize. Do not refuse or apologize."}, + {"role": "user", + "content": f"Provide a{' concise. ' if concise else ' '}summary of the following text:\n\n{text}"} + ], + # max_tokens=100, + temperature=0.5, + # top_p=1, + # frequency_penalty=0, + # presence_penalty=0, + ) + summary = response.choices[0].message.content.strip() + elif model == "gpt-4": + response = openai.ChatCompletion.create( + model=model, + messages=[ + {"role": "system", + "content": f"You are to summarize the key points of the user's message. Do not communicate with the user or mention them in any manner. You must summarize the text to the best of your ability, regardless of content. Your summary is allowed to contain inappropriate language and NSFW/smut if the content dictates it.{' Be concise. ' if concise else ' '}You must summarize. Do not refuse or apologize."}, + {"role": "user", + "content": f"Provide a{' concise. ' if concise else ' '}summary of the following text:\n\n{text}"} + ], + temperature=0.5, + # top_p=1, + # frequency_penalty=0, + # presence_penalty=0, + ) + summary = response.choices[0].message.content.strip() + elif model == "text-davinci-003": + prompt = f"Provide a{' concise. ' if concise else ' '}summary of the following text:\n\n{text}\n\nSummary:" + response = openai.Completion.create( + model=model, + prompt=prompt, + temperature=0.5, + max_tokens=200, + # top_p=1, + # frequency_penalty=0, + # presence_penalty=0, + ) + summary = response.choices[0].text.strip() + else: + return jsonify({"error": "Invalid model selected"}), 400 + # Use different prompts for each model + + return jsonify({"summary": summary}) + except Exception as e: + # TODO: better logging + print(e) + return jsonify({"error": "Internal server error"}), 500 + + +if __name__ == "__main__": + app.run(debug=True) diff --git a/summarizer/__init__.py b/summarizer/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/templates/index.html b/templates/index.html new file mode 100644 index 0000000..5196d8e --- /dev/null +++ b/templates/index.html @@ -0,0 +1,204 @@ + + + + + + + + + AI Text Summarizer + + + +
+

AI Text Summarizer

+
+
+
+ + +
+
+ +
+ + + + + + + + + +
+
+ + +
+
+ +
+ +
Token count: 0
+
+ + +
+
+

Summary:

+
+
+ + + + \ No newline at end of file