first commit

Author: TheRanomial

.gitignore

@@ -0,0 +1,3 @@
+
+
+.env

__pycache__/constants.cpython-312.pyc

@@ -0,0 +1,37 @@
+from fastapi import APIRouter
+import base64
+from io import BytesIO
+from schemareq import ImageData
+from apps.calculator.utils2 import understand_image
+from apps.calculator.utils3 import generate_image
+from PIL import Image
+
+router=APIRouter()
+
+@router.post('')
+async def run(imageData:ImageData):
+    image_data=base64.b64decode(imageData.image.split(",")[1])
+    image_bytes=BytesIO(image_data)
+    image=Image.open(image_bytes)
+    responses=understand_image(img=image,dict_of_vars=imageData.dict_of_vars)
+    
+    data=[]
+    for response in responses:
+        data.append(response)
+        
+    ##image_details=str(response)
+
+    print('response in route: ', response.get("Attributes"))
+    attributes=response.get("Attributes")
+    primary_object=response.get("Primary Object")
+    context=response.get("Context")
+    style_notes=response.get("Style Notes")
+    
+    generated_image=generate_image(attributes,primary_object,context,style_notes);
+    
+    buffer = BytesIO()
+    generated_image.save(buffer, format="PNG")
+    buffer.seek(0)
+    base64_image = base64.b64encode(buffer.getvalue()).decode("utf-8")
+    
+    return {"message": "Image processed", "data": data, "status": "success","image": f"data:image/png;base64,{base64_image}",}

__pycache__/main.cpython-312.pyc

@@ -0,0 +1,23 @@
+from fastapi import APIRouter
+import base64
+from io import BytesIO
+from schemareq import ImageData
+from apps.calculator.utils import analyze_image
+from PIL import Image
+
+router=APIRouter()
+
+@router.post('')
+async def run(imageData:ImageData):
+    image_data=base64.b64decode(imageData.image.split(",")[1])
+    image_bytes=BytesIO(image_data)
+    image=Image.open(image_bytes)
+    responses=analyze_image(img=image,dict_of_vars=imageData.dict_of_vars)
+    
+    data=[]
+    for response in responses:
+        data.append(response)
+    print('response in route: ', response)
+    return {"message": "Image processed", "data": data, "status": "success"}
+    
+    

__pycache__/schemareq.cpython-312.pyc

@@ -0,0 +1,90 @@
+import google.generativeai as genai
+import os
+from dotenv import load_dotenv
+import ast
+import json
+from PIL import Image
+
+load_dotenv()
+
+gemini_api_key=os.getenv("GEMINI_API_KEY")
+
+genai.configure(api_key=gemini_api_key)
+
+def analyze_image(img: Image, dict_of_vars: dict):
+    model = genai.GenerativeModel(model_name="gemini-1.5-flash")
+    dict_of_vars_str = json.dumps(dict_of_vars, ensure_ascii=False)
+    prompt = (
+    f"You are an advanced mathematical problem solver tasked with analyzing and solving mathematical expressions from images. "
+
+    f"Mathematical Expression Rules:\\n"
+    f"Follow the PEMDAS rule for solving all mathematical expressions:\\n"
+    f"- P: Parentheses (highest priority)\\n"
+    f"- E: Exponents\\n"
+    f"- M/D: Multiplication and Division (left to right)\\n"
+    f"- A/S: Addition and Subtraction (left to right)\\n\\n"
+
+    f"Examples with Detailed Solutions:\\n"
+    f"1. Expression: 2 + 3 * 4\\n"
+    f"   Step-by-step solution:\\n"
+    f"   - First multiply: 3 * 4 = 12 (multiplication before addition)\\n"
+    f"   - Then add: 2 + 12 = 14\\n"
+    f"   Final answer: 14\\n\\n"
+    
+    f"2. Expression: 2 + 3 + 5 * 4 - 8 / 2\\n"
+    f"   Step-by-step solution:\\n"
+    f"   - First handle multiplication: 5 * 4 = 20\\n"
+    f"   - Then division: 8 / 2 = 4\\n"
+    f"   - Finally, perform addition and subtraction left to right: 2 + 3 + 20 - 4 = 21\\n"
+    f"   Final answer: 21\\n\\n"
+    
+    
+    f"Problem Types and Return Formats:\\n\\n"
+    
+    f"1. Simple Mathematical Expressions\\n"
+    f"   Input example: 2 + 2, 3 * 4, etc.\\n"
+    f"   Return format: [{{'expr': 'given_expression', 'result': calculated_answer}}]\\n"
+    f"   Example: [{{'expr': '2 + 2', 'result': 4}}]\\n\\n"
+    
+    f"2. Systems of Equations\\n"
+    f"   Input example: x^2 + 2x + 1 = 0, 3y + 4x = 0\\n"
+    f"   Return format: [{{'expr': 'x', 'result': value1, 'assign': True}}, "
+    f"                   {{'expr': 'y', 'result': value2, 'assign': True}}]\\n\\n"
+    
+    f"3. Variable Assignments\\n"
+    f"   Input example: x = 4, y = 5\\n"
+    f"   Return format: [{{'expr': 'x', 'result': 4, 'assign': True}}]\\n\\n"
+    
+    f"4. Graphical Math Problems\\n"
+    f"   Description: Problems represented through drawings (collisions, triangles, etc.)\\n"
+    f"   Special instruction: Consider color variations in drawings\\n"
+    f"   Return format: [{{'expr': 'problem_description', 'result': calculated_answer}}]\\n\\n"
+    
+    f"5. Abstract Concept Detection\\n"
+    f"   Description: Identifying abstract concepts from drawings\\n"
+    f"   Return format: [{{'expr': 'drawing_explanation', 'result': 'abstract_concept'}}]\\n\\n"
+    
+    f"Variable Resolution Rules:\\n"
+    f"- Use this dictionary for pre-assigned variables: {dict_of_vars_str}\\n"
+    f"- Replace all variables with their corresponding values before computation\\n"
+    f"- If a variable is not found in the dictionary, treat it as an unknown to be solved\\n\\n"
+    
+    f"DO NOT USE BACKTICKS OR MARKDOWN FORMATTING. "
+    f"PROPERLY QUOTE THE KEYS AND VALUES IN THE DICTIONARY FOR EASIER PARSING WITH Python's ast.literal_eval."
+    )
+    
+    response = model.generate_content([prompt, img])
+    print(response.text)
+    answers = []
+    try:
+        answers = ast.literal_eval(response.text)
+    except Exception as e:
+        print(f"Error in parsing response from Gemini API: {e}")
+    print('returned answer ', answers)
+    for answer in answers:
+        if 'assign' in answer:
+            answer['assign'] = True
+        else:
+            answer['assign'] = False
+    return answers
+

apps/calculator/__pycache__/image.cpython-312.pyc

@@ -0,0 +1,54 @@
+import google.generativeai as genai
+import os
+from dotenv import load_dotenv
+import ast
+from PIL import Image
+
+load_dotenv()
+
+gemini_api_key=os.getenv("GEMINI_API_KEY")
+
+genai.configure(api_key=gemini_api_key)
+
+def understand_image(img: Image, dict_of_vars: dict):
+    model = genai.GenerativeModel(model_name="gemini-1.5-flash")
+    prompt = (
+    f"You are an AI model designed to analyze and interpret drawings provided by users. "
+    
+    f"Primary Object Identification:\\n"
+    f"Identify the main object or theme depicted in the drawing. Be creative but stay grounded in the details of the image.\\n\\n"
+    
+    f"Attributes Description:\\n"
+    f"Describe key characteristics of the drawing such as shape, size, texture, or patterns. Consider any distinct visual features that stand out.\\n\\n"
+    
+    f"Contextual Details:\\n"
+    f"Add possible context based on the drawing. This can include the environment, setting, or mood of the image. If there are interactions or relationships between objects, mention them.\\n\\n"
+    
+    f"Style Notes:\\n"
+    f"Describe the style of the drawing, if any. Is it sketchy, detailed, minimalist, or abstract? Also, note the perspective (e.g., top-down view, side view, etc.) and any specific artistic techniques used.\\n\\n"
+    
+    f"Return Format: [{{'Primary Object': 'description', 'Attributes': 'description', 'Context': 'description', 'Style Notes': 'description'}}]\\n"
+    f"Example: [{{'Primary Object': 'tree', 'Attributes': 'a simple tree with a trunk and leafy branches', 'Context': 'the tree is drawn against a plain background', 'Style Notes': 'sketched with basic lines and no shading'}}]\\n\\n"
+    
+    f"DO NOT USE BACKTICKS OR MARKDOWN FORMATTING. "
+    f"PROPERLY QUOTE THE KEYS AND VALUES IN THE DICTIONARY FOR EASIER PARSING WITH Python's ast.literal_eval."
+    )
+    
+    response = model.generate_content([prompt, img])
+    
+    answers = []
+    try:
+        answers = ast.literal_eval(response.text)
+    except Exception as e:
+        print(f"Error in parsing response from Gemini API: {e}")
+   
+    for answer in answers:
+        if 'assign' in answer:
+            answer['assign'] = True
+        else:
+            answer['assign'] = False
+       
+    return answers;
+    
+
+    

apps/calculator/__pycache__/route.cpython-312.pyc

@@ -0,0 +1,61 @@
+import os
+from PIL import Image
+from io import BytesIO
+import requests
+from urllib.parse import quote
+from dotenv import load_dotenv
+import google.generativeai as genai
+
+load_dotenv()
+
+gemini_api_key = os.getenv("GEMINI_API_KEY")
+genai.configure(api_key=gemini_api_key)
+
+def generate_image(attributes: str, primary_object: str, context: str, style_notes: str):
+    final_prompt = f"Create an image of a {attributes} {primary_object} {context}. "
+    if style_notes:
+        final_prompt += f"Style: {style_notes}. "
+
+    enhanced_prompt = (
+        f"{final_prompt}, "
+        "high resolution, highly detailed, sharp focus, "
+        "professional photography, cinematic lighting, "
+        "8k uhd, ray tracing, ambient lighting"
+    )
+
+    negative_prompt = (
+        "blurry, low quality, low resolution, "
+        "watermark, signature, oversaturated, "
+        "distorted, deformed, pixelated"
+    )
+
+    base_url = "https://image.pollinations.ai/prompt"
+    encoded_prompt = quote(enhanced_prompt)
+    url = f"{base_url}/{encoded_prompt}"
+
+    params = {
+        "width": 1024,
+        "height": 1024,
+        "model": "flux-pro",
+        "sampler": "DPM++ SDE Karras",
+        "enhance": True,
+        "nologo": True,
+        "negative_prompt": negative_prompt,
+        "upscale": True,
+        "upscale_amount": "2",
+    }
+
+    try:
+        response = requests.get(url, params=params, stream=True)
+        response.raise_for_status()
+    except requests.exceptions.RequestException as e:
+        print(f"Request failed: {e}")
+        return None
+    if response.status_code == 200:
+        image = Image.open(BytesIO(response.content))
+        return image
+    else:
+        print(f"Error: {response.status_code} - {response.text}")
+        return None
+
+

apps/calculator/__pycache__/utils.cpython-312.pyc

@@ -0,0 +1,9 @@
+import os
+from dotenv import load_dotenv
+load_dotenv()
+
+SERVER_URL = 'localhost'
+PORT = '8900'
+ENV = 'dev'
+
+GEMINI_API_KEY =os.getenv("GEMINI_API_KEY")

apps/calculator/__pycache__/utils2.cpython-312.pyc

@@ -0,0 +1,33 @@
+from contextlib import asynccontextmanager
+from fastapi import FastAPI
+from fastapi.middleware.cors import CORSMiddleware
+import uvicorn
+from constants import SERVER_URL,PORT,ENV
+from apps.calculator.route import router as calculator_router
+from apps.calculator.image import router as image_router
+
+@asynccontextmanager
+async def lifespan(app:FastAPI):
+    yield
+    
+
+app=FastAPI(lifespan=lifespan)
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+@app.get("/")
+async def root():
+    return {"server is running good"}
+
+app.include_router(calculator_router,prefix="/calculate",tags=["calculate"])
+app.include_router(image_router,prefix="/generate",tags=["generate"])
+
+if __name__ == "__main__":
+    import uvicorn
+    uvicorn.run(app, host=SERVER_URL, port=int(PORT),reload=(ENV=="dev"))

apps/calculator/__pycache__/utils3.cpython-312.pyc

@@ -0,0 +1,7 @@
+google-generativeai
+uvicorn
+pydantic
+fastapi
+Pillow
+python-dotenv
+requests

apps/calculator/image.py

@@ -0,0 +1,5 @@
+from pydantic import BaseModel
+
+class ImageData(BaseModel):
+    image:str
+    dict_of_vars:dict