Update .env.example and requirements.txt, delete unused files

src/.env.example

@@ -1 +1,2 @@
 OPENAI_API_KEY=""
+AZURE = ""

src/app/api/models.py

@@ -4,7 +4,7 @@ class Product(models.Model):
     barcode = models.CharField(max_length=20)
     brand = models.CharField(max_length=100)
     sub_brand = models.CharField(max_length=100, blank=True, null=True)
-    manufacturer = models.CharField(max_length=200)
+    manufactured_by = models.CharField(max_length=200)
     product_name = models.CharField(max_length=200)
     weight = models.FloatField()
     variant = models.CharField(max_length=100, blank=True, null=True)
@@ -14,6 +14,30 @@ class Product(models.Model):
     child_category = models.CharField(max_length=100)
     sub_child_category = models.CharField(max_length=100, blank=True, null=True)
     images_paths = models.CharField(max_length=3000, blank=True, null=True) # Comma separated paths
+    description = models.TextField(max_length=3000, blank=True, null=True)
+    quantity = models.IntegerField(null=True, blank=True)
+    mrp = models.CharField(max_length=100, blank=True, null=True)
 
-    def __str__(self):
-        return self.product_name
+    def _str_(self):
+        return self.product_name
+
+
+class Database(models.Model):
+    barcode = models.CharField(max_length=20)
+    brand = models.CharField(max_length=100)
+    sub_brand = models.CharField(max_length=100, blank=True, null=True)
+    manufactured_by = models.CharField(max_length=200)
+    product_name = models.CharField(max_length=200)
+    weight = models.FloatField()
+    variant = models.CharField(max_length=100, blank=True, null=True)
+    net_content = models.CharField(max_length=100, blank=True, null=True)
+    price = models.DecimalField(max_digits=10, decimal_places=2)
+    parent_category = models.CharField(max_length=100)
+    child_category = models.CharField(max_length=100)
+    sub_child_category = models.CharField(max_length=100, blank=True, null=True)
+    images_paths = models.CharField(max_length=3000, blank=True, null=True) # Comma separated paths
+    description = models.TextField(max_length=3000, blank=True, null=True)
+    quantity = models.IntegerField(null=True, blank=True)
+    promotion_on_the_pack = models.CharField(max_length=100, blank=True, null=True)
+    type_of_packaging = models.CharField(max_length=100, blank=True, null=True)
+    mrp = models.CharField(max_length=100, blank=True, null=True)

src/module/audio_text.py

@@ -1,49 +0,0 @@
-# from whisper_jax import FlaxWhisperPipline
-# import jax.numpy as jnp
-import whisper
-print(whisper.__file__)
-from openai import OpenAI
-from module.config import OPENAI_API_KEY
-import os
-
-client = OpenAI()
-os.environ['OPENAI_API_KEY'] = OPENAI_API_KEY
-
-
-def whisper_pipeline_tpu(audio):
-    pipeline = FlaxWhisperPipline("openai/whisper-large-v3", dtype=jnp.bfloat16, batch_size=16)
-    text = pipeline(audio)
-    return text
-
-
-     
-def whisper_pipeline(audio_path):
-    model = whisper.load_model("medium")
-    # load audio and pad/trim it to fit 30 seconds
-    audio = whisper.load_audio(audio_path)
-    audio = whisper.pad_or_trim(audio)
-    # make log-Mel spectrogram and move to the same device as the model
-    mel = whisper.log_mel_spectrogram(audio).to(model.device)
-    # detect the spoken language
-    _, probs = model.detect_language(mel)
-    print(f"Detected language: {max(probs, key=probs.get)}")
-    # decode the audio
-    options = whisper.DecodingOptions()
-    result = whisper.decode(model, mel, options)
-    # print the recognized text
-    print(result.text)
-    return result.text
-
-
-
-
-
-def whisper_openai(audio_path):
-   audio_file= open(audio_path, "rb")
-   transcript = client.audio.transcriptions.create(
-    model="whisper-1", 
-    file=audio_file
-   )
-   return transcript
-
-whisper_pipeline()

src/module/image.ipynb

@@ -1,44 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": 1,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import cv2\n",
-    "import os\n",
-    "import matplotlib.pyplot as plt \n",
-    "import numpy as np "
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": []
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "catlog",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.10.13"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 2
-}

src/module/image_enhance.py

@@ -1,111 +0,0 @@
-import cv2
-import os
-from config import file_Directory
-import numpy as np
-from PIL import Image
-
-class Image_Enhance():
-
-    def __init__(self, image_path) -> None:
-        self.image_path = image_path
-    
-    def brightness_Adjust(self):
-        # Load the image 
-        image = cv2.imread(self.image_path) 
-        #Plot the original image 
-        alpha = -1.1
-        # control brightness by 50 
-        beta = 70
-        image2 = cv2.convertScaleAbs(image, alpha=alpha, beta=beta) 
-        #Save the image 
-        # imagepth = os.path.join(os.path.dirname(self.image_path), 'Brightness & contrast.jpg')
-        imagepth = os.path.join(file_Directory, 'Brightness & contrast.jpg')
-        cv2.imwrite(imagepth, image2) 
-        return imagepth
-    
-    def remove_flash(self, imagepth):
-        image = cv2.imread(imagepth)
-        # cv2.cvtColor is applied over the 
-        gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
-
-        # Apply adaptive thresholding to segment the text
-        thresh = cv2.adaptiveThreshold(gray, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY_INV, 11, 4)
-
-        # Apply Gaussian blur to the grayscale image to reduce noise
-        blurred = cv2.GaussianBlur(gray, (5, 5), 0)
-
-        # Threshold the blurred image to create a binary mask for the flashlight glare
-        _, mask = cv2.threshold(blurred, 240, 255, cv2.THRESH_BINARY_INV)
-
-        # Combine the text and glare masks
-        mask = cv2.bitwise_or(mask, thresh)
-
-        # Apply morphological closing to further remove small areas of glare
-        kernel = np.ones((5,5),np.uint8)
-        mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, kernel)
-
-        # Apply the mask to the original image to remove flashlight glare
-        result = cv2.bitwise_and(image, image, mask=mask)
-
-        cv2.imwrite(os.path.join(file_Directory, 'remove_flash.jpg'), result)
-            
-    def sharpen(self, imagepth):
-        image = cv2.imread(imagepth) 
-        # Create the sharpening kernel 
-        kernel = np.array([[0, -1, 0], [-1, 5, -1], [0, -1, 0]]) 
-        # Sharpen the image 
-        sharpened_image = cv2.filter2D(image, -1, kernel) 
-        #Save the image
-        imagepath = os.path.join(file_Directory, 'sharpened_image.jpg') 
-        cv2.imwrite(imagepath, sharpened_image) 
-        return imagepath
-        
-
-    def lapacian_sharpen(self, imagepth):
-        #Load the image 
-        image = cv2.imread(imagepth) 
-        
-        # Sharpen the image using the Laplacian operator 
-        sharpened_image2 = cv2.Laplacian(image, cv2.CV_64F) 
-        imagepath = os.path.join(file_Directory, 'Laplacian_sharpened_image.jpg') 
-        #Save the image 
-        cv2.imwrite(imagepath, sharpened_image2) 
-
-    def removing_noise(self, imagepth):
-        # Load the image 
-        image = cv2.imread(imagepth) 
-        # Remove noise using a median filter 
-        filtered_image = cv2.medianBlur(image, 1) 
-        imagepath = os.path.join(file_Directory, 'Median Blur.jpg') 
-        #Save the image 
-        cv2.imwrite(imagepath, filtered_image) 
-
-        return imagepath
-
-    
-    def enhance_color(self, imagepth):
-        # Load the image 
-        image = cv2.imread(imagepth) 
-        image = cv2.cvtColor(image, cv2.COLOR_RGB2HSV) 
-        
-        # Adjust the hue, saturation, and value of the image 
-        # Adjusts the hue by multiplying it by 0.7 
-        image[:, :, 0] = image[:, :, 0] * 0.7
-        # Adjusts the saturation by multiplying it by 1.5 
-        image[:, :, 1] = image[:, :, 1] * 1.5
-        # Adjusts the value by multiplying it by 0.5 
-        image[:, :, 2] = image[:, :, 2] * 0.5
-        
-        image2 = cv2.cvtColor(image, cv2.COLOR_HSV2BGR) 
-        imagepath = os.path.join(file_Directory, 'enhanced coloured.jpg') 
-        #Save the image 
-        cv2.imwrite(imagepath, image2) 
-       
-
-obj = Image_Enhance(r"data/Catalog Digitization/ONDC Test Data _ Images/Product Images/Bru_Instant_Coffee_Powder.png")
-pth = obj.brightness_Adjust()
-sharpen = obj.sharpen(pth)
-lapacian_sharpen = obj.lapacian_sharpen(sharpen)
-noise = obj.removing_noise(sharpen)
-obj.enhance_color(noise)
-obj.remove_flash(sharpen)

src/module/llm_vision.py

@@ -1,59 +0,0 @@
-import base64
-import requests
-from config import OPENAI_API
-import os
-
-OPENAI_API = os.getenv("OPENAI_API")
-
-""" 
-openai_vision = OpenAIVision(api_key)
-image_path = "path_to_your_image.jpg"
-prompt = ""
-response = openai_vision.get_image_description(prompt,image_path)
-"""
-
-class OpenAIVision:
-    def __init__(self):
-        self.api_key = OPENAI_API
-        self.base_url = "https://api.openai.com/v1/chat/completions"
-
-    def __encode_image(self, image_path):
-        with open(image_path, "rb") as image_file:
-            return base64.b64encode(image_file.read()).decode('utf-8')
-
-    def get_image_description(self, image_path, prompt):
-        base64_image = self.__encode_image(image_path)
-
-        headers = {
-            "Content-Type": "application/json",
-            "Authorization": f"Bearer {self.api_key}"
-        }
-
-        payload = {
-            "model": "gpt-4-vision-preview",
-            "temperature": 0.0,
-            "messages": [
-                {
-                    "role": "user",
-                    "content": [
-                        {
-                            "type": "text",
-                            "text": prompt,
-                        },
-                        {
-                            "type": "image_url",
-                            "image_url": {
-                                "url": f"data:image/jpeg;base64,{base64_image}"
-                            }
-                        }
-                        
-                    ]
-                  
-                }
-            ],
-            "max_tokens": 1000,
-            
-        }
-
-        response = requests.post(self.base_url, headers=headers, json=payload)
-        return response.json()

src/module/prompts/base.py

@@ -1,34 +0,0 @@
-from textwrap import dedent
-
-base_prompt = dedent("""
-        ### Instruction:
-        product description starts here
-
-        {text}
-
-        product description ends here
-
-        this is the categorys list ['BEVERAGES', 'SNACKS & BRANDED FOODS', 'NOT FOUND', 'EGGS, MEAT & FISH', 'FOODGRAINS, OIL & MASALA', 'PERSONAL CARE', 'CLEANING & HOUSEHOLD', 'FRUITS & VEGETABLES', 'BAKERY, CAKES & DAIRY', 'MAKEUP', 'BABY CARE', 'PET FOOD & ACCESSORIES', 'NON FMCG', 'ALCOHOL & TOBACCO', 'WELLNESS', 'EVERYDAY MEDICINE-NEW', 'EXCERCISE & FITNESS', 'ALCOHOLIC BEVERAGES'].
-
-        Analyse data from the above product description to give me the following details in JSON format:
-        ( return "null" where you don't have a answer)
-        
-        "brand": "sample_brand",
-        "mrp": "The price might start with MRP or Rs.",
-        "unit": "per pack",
-        "Quantity": 1,  ##num of products visible
-        "parent_category": "from the above given list",
-        "ingredients": ["ingredient1", "ingredient2", "ingredient3"],
-        "calorie_count": "Would be in numbers",
-        "marketed_by": "sample_marketer",
-        "manufactured_by": "sample_manufacturer",
-        "manufactured_in_country": "Country XYZ",
-        "type_of_packaging": "Box",
-        "promotion_on_the_pack": "if any",
-        "type_of_product": "give this your understanding",
-        "pack_of_or_no_of_units": "No. of Units"
-     
-
-        Analyse data from the above product description to give me the following details in JSON format:
-        Only return the output in the required json format.
-        """)

src/module/vectorsearch.py

@@ -1,38 +0,0 @@
-import os
-from config import OPENAI_API_KEY, file_Directory
-from langchain_community.document_loaders.csv_loader import CSVLoader
-from langchain_openai import OpenAIEmbeddings
-from langchain.text_splitter import CharacterTextSplitter
-from langchain_community.vectorstores import Chroma
-import pandas as pd
-
-os.environ['OPENAI_API_KEY'] = OPENAI_API_KEY
-
-
-# df = pd.read_excel(r"/home/vrush/Catalog-Digitization-/src/module/data/Catalog Digitization/ONDC Test Data _ Images/ONDCSampleData.xlsx")
-# df_new = pd.DataFrame(columns=["id", "name"])
-# df_new =  df['name']
-# df_new.to_csv(r"data/data.csv", index=False)
-
-def create_vector():
-    loader = CSVLoader(file_path="data/data.csv")
-    docs = loader.load()
-    text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
-    documents = text_splitter.split_documents(docs)
-    db_path = os.path.join(file_Directory,"vectorstore")
-    embeddings = OpenAIEmbeddings()
-    os.makedirs(db_path, exist_ok=True)      
-    Chroma.from_documents(docs, embeddings, persist_directory= db_path)
-    
-def search(query):
-    embeddings = OpenAIEmbeddings()
-    db_path = os.path.join(file_Directory,"vectorstore")
-    db = Chroma(persist_directory= db_path, embedding_function= embeddings)
-    embedding_vector = OpenAIEmbeddings().embed_query(query)
-    docs = db.similarity_search_by_vector(embedding_vector)
-    print(docs[0].page_content)
-
-
-if __name__ == "__main__":
-    create_vector()
-    search("Choco Creme Wafers")

src/requirements.txt

@@ -3,4 +3,8 @@ langchain==0.1.6
 python-decouple==3.4
 pandas
 azure-ai-formrecognizer
-easyocr
+easyocr
+langchain
+chromadb
+langchain_openai
+unstructured