import gradio as gr import requests from PIL import Image from io import BytesIO import tempfile from gtts import gTTS import re import graphviz import os # Load the API key from environment variables for security API_KEY = '715d7f1ce56d4c1abb3a803e77ffae87' # Define API endpoints IMAGE_API_URL = 'https://api.aimlapi.com/images/generations' CHAT_API_URL = 'https://api.aimlapi.com/chat/completions' # List of available chat models CHAT_MODELS = [ "meta-llama/Meta-Llama-3-8B-Instruct-Lite", "meta-llama/Meta-Llama-3-70B-Instruct-Lite", "meta-llama/Meta-Llama-3-70B-Instruct-Turbo", "meta-llama/Meta-Llama-3-8B-Instruct-Turbo", "gpt-4o" ] # Load supported languages from a file def load_languages(file_path='languages.txt'): languages = {} try: with open(file_path, 'r', encoding='utf-8') as file: for line in file: if line.strip(): language, code = line.strip().split(': ') languages[language] = code except FileNotFoundError: print(f"Error: {file_path} not found.") return languages languages = load_languages() def get_answer_content(language_name, question, model_name, category, max_chars, max_lines): language_code = languages.get(language_name, 'en') headers = { 'Authorization': f'Bearer {API_KEY}', 'Content-Type': 'application/json' } data = { "model": model_name, "messages": [ { "role": "user", "content": f"Respond in {language_name} for category '{category}': {question}" } ], "max_tokens": 1500, "stream": False } try: response = requests.post(CHAT_API_URL, headers=headers, json=data) response.raise_for_status() answer_content = response.json()['choices'][0]['message']['content'] if category in ["Documentation", "Research"]: answer_content = answer_content[:1500] # Truncate to max_chars if max_chars: answer_content = answer_content[:int(max_chars)] # Ensure the output ends with a complete sentence if max_chars: truncated_length = int(max_chars) if truncated_length < len(answer_content): # Find the last sentence-ending punctuation within the truncated length last_punctuation_index = max( answer_content.rfind(p) for p in ".!?" ) if last_punctuation_index > -1 and last_punctuation_index <= truncated_length: answer_content = answer_content[:last_punctuation_index + 1] else: # If no punctuation is found or it's outside the limit, truncate at the limit answer_content = answer_content[:truncated_length] # Limit by max_lines if specified if max_lines: answer_content = "\n".join(answer_content.splitlines()[:int(max_lines)]) # Remove unwanted introductory lines lines = answer_content.splitlines() filtered_lines = [line for line in lines if not line.lower().startswith("here's a joke about")] filtered_content = "\n".join(filtered_lines) return filtered_content except requests.RequestException as e: return f"An error occurred: {e}" def preprocess_text(text): return re.sub(r'[^\w\s,.!?]', '', text) def text_to_speech_online(text, lang='en'): try: cleaned_text = preprocess_text(text) tts = gTTS(text=cleaned_text, lang=lang, slow=False) with tempfile.NamedTemporaryFile(delete=False, suffix='.mp3') as temp_file: tts.save(temp_file.name) return temp_file.name except Exception as e: print(f"Text-to-speech failed: {e}") return None def generate_image(prompt, model_name): headers = {"Authorization": f"Bearer {API_KEY}"} payload = {"prompt": prompt, "model": model_name} try: response = requests.post(IMAGE_API_URL, headers=headers, json=payload) response.raise_for_status() output = response.json() if "images" in output and output["images"]: image_url = output["images"][0]["url"] img_data = requests.get(image_url).content image = Image.open(BytesIO(img_data)) return image else: print("Unexpected response structure:", output) return Image.new('RGB', (512, 512), color=(255, 0, 0)) except requests.exceptions.RequestException as e: print(f"An error occurred: {e}") return Image.new('RGB', (512, 512), color=(255, 0, 0)) except Exception as e: print(f"An unexpected error occurred: {e}") return Image.new('RGB', (512, 512), color=(255, 0, 0)) def wrap_text(text, width=30): words = text.split() lines, current_line, current_length = [], [], 0 for word in words: if current_length + len(word) <= width: current_line.append(word) current_length += len(word) + 1 else: lines.append(" ".join(current_line)) current_line = [word] current_length = len(word) + 1 lines.append(" ".join(current_line)) return "\n".join(lines) def generate_workflow_diagram(steps): dot = graphviz.Digraph(format='png') dot.attr(rankdir='TB', size='10,10', nodesep='0.5', ranksep='0.5', dpi='300') steps = steps.strip().split("\n") if not steps or steps == [""]: return None for i, step in enumerate(steps): step = wrap_text(step.strip(), width=30) if step: dot.node(str(i), step, shape='box', width='2.0', height='0.5', fontsize='12') if i > 0: dot.edge(str(i - 1), str(i)) with tempfile.NamedTemporaryFile(delete=False, suffix='.png') as temp_file: dot.render(temp_file.name) return temp_file.name + '.png' def on_button_click(language_name, question, model_name, category, max_chars, max_lines): if not question.strip(): return "Please enter a question.", None if category == "default": category = "Post" answer = get_answer_content(language_name, question, model_name, category, max_chars, max_lines) audio_file = text_to_speech_online(answer, languages.get(language_name, 'en')) return f"You: {question}\n\nAI MINDS:\n\n{answer}", audio_file def on_image_button_click(prompt, model_name): return generate_image(prompt, model_name) def on_workflow_button_click(steps): return generate_workflow_diagram(steps) def clear_all(): return None, None, None, None, None # Define Gradio Interface with gr.Blocks() as demo: gr.Markdown("# AI_MINDS CHATPLUS") with gr.Tabs(): with gr.Tab("Chat"): gr.Markdown("## Chat Section") with gr.Row(): language_dropdown = gr.Dropdown(choices=list(languages.keys()), label="Select Language", value="English") model_dropdown = gr.Dropdown(choices=CHAT_MODELS, label="Select Chat Model", value="meta-llama/Meta-Llama-3-70B-Instruct-Turbo") category_dropdown = gr.Dropdown(choices=["default","Post", "Documentation", "Research", "Generation"], label="Select Category", value="default") max_chars_input = gr.Number(label="Max Characters (Optional)", value=None, step=1, precision=0) max_lines_input = gr.Number(label="Max Lines (Optional)", value=None, step=1, precision=0) with gr.Row(): with gr.Column(scale=1): question_input = gr.Textbox(label="Your Question", placeholder='Ask a question...', lines=2) generate_button = gr.Button("Ask") small_audio_output = gr.Audio(label="Voice Output", type="filepath", visible=True, interactive=False) clear_button = gr.Button("Clear") with gr.Column(scale=2): content_output = gr.Markdown(label="Chat Output") with gr.Tab("Image"): gr.Markdown("## Image Generation Section") image_prompt = gr.Textbox(label="Image Prompt", placeholder='Enter an image prompt...') image_model_dropdown = gr.Dropdown(choices=["flux-realism", "stable-diffusion-v3-medium"], label="Select Image Model", value="flux-realism") generated_image = gr.Image(label="Generated Image", type="pil") image_generate_button = gr.Button("Generate Image") with gr.Tab("Flowchart"): gr.Markdown("## Workflow Diagram Generator") workflow_input = gr.Textbox(lines=10, placeholder="Enter workflow steps, one per line.", label="Workflow Steps \n\n it is giving error in flowChart generation because of some dependencies issues in Hugging Face Hosting \n\n please check in colab notebook from GitHub respository for workflow perocess.\n Link is here \n\n https://github.com/shahid9455/AI_MINDS_GPTPLUS") generate_workflow_button = gr.Button("Generate Diagram") diagram_output = gr.Image(label="Generated Workflow Diagram") # Define button actions generate_button.click(on_button_click, [language_dropdown, question_input, model_dropdown, category_dropdown, max_chars_input, max_lines_input], [content_output, small_audio_output]) image_generate_button.click(on_image_button_click, [image_prompt, image_model_dropdown], [generated_image]) generate_workflow_button.click(on_workflow_button_click, [workflow_input], [diagram_output]) clear_button.click(fn=clear_all, inputs=[], outputs=[question_input, content_output, generated_image, diagram_output, small_audio_output]) # Launch the Gradio app if __name__ == "__main__": demo.launch()