Upload 6 files

animation.ipynb

@@ -0,0 +1,66 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "import json \n",
+    "import requests \n",
+    "  \n",
+    "import streamlit as st \n",
+    "from streamlit_lottie import st_lottie \n",
+    "import os\n",
+    "folder_path = \"lotte_animation_saver/\"\n",
+    "\n",
+    "# Get the list of all files in the folder\n",
+    "file_names = os.listdir(folder_path)\n",
+    "max_file_name=max([int(i.split(\"_\")[-1][:-5]) for i in file_names])  \n",
+    "url = requests.get( \n",
+    "    \"https://lottie.host/34b4b005-c8c9-4d57-a0fd-0e9ed3b49835/nCkgflSg9J.json\") \n",
+    "# Creating a blank dictionary to store JSON file, \n",
+    "# as their structure is similar to Python Dictionary \n",
+    "url_json = dict() \n",
+    "  \n",
+    "if url.status_code == 200: \n",
+    "\n",
+    "    url_json = url.json() \n",
+    "    with open(\"lotte_animation_saver/animation_\"+str(max_file_name+1)+\".json\",\"w\") as read:\n",
+    "        json.dump(url_json,read,indent=2)\n",
+    "\n",
+    "else: \n",
+    "    print(\"Error in the URL\") "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "env",
+   "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": "3.12.4"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

app.py

@@ -0,0 +1,440 @@
+import pandas as pd
+from PIL import Image
+import streamlit as st
+import cv2
+from streamlit_drawable_canvas import st_canvas
+import torch
+from diffusers import AutoPipelineForInpainting
+import numpy as np
+from streamlit_image_select import image_select
+import os
+import requests 
+from streamlit_navigation_bar import st_navbar
+from langchain_community.llms import Ollama
+import base64
+from io import BytesIO
+from PIL import Image, ImageDraw
+from streamlit_lottie import st_lottie 
+from streamlit_option_menu import option_menu
+import json
+from transformers import pipeline
+import streamlit as st
+from streamlit_modal import Modal
+import streamlit.components.v1 as components
+from datetime import datetime
+
+def image_to_base64(image_path):
+    with open(image_path, "rb") as img_file:
+        return base64.b64encode(img_file.read()).decode()
+    
+
+@st.cache_resource
+def load_model():
+    pipeline_ = AutoPipelineForInpainting.from_pretrained("kandinsky-community/kandinsky-2-2-decoder-inpaint", torch_dtype=torch.float16).to("cuda")
+    return pipeline_
+
+# @st.cache_resource
+def prompt_improvment(pre_prompt):
+
+    llm = Ollama(model="llama3:latest",num_ctx=1000)
+    enhancement="Please use details from the prompt mentioned above, focusing only what user is thinking with the prompt and also add 8k resolution. Its a request only provide image description and brief prompt no other text."
+    prompt = pre_prompt+"\n"+enhancement
+    # result = llm.invoke(prompt)
+    return llm.stream(prompt)
+def numpy_to_list(array):
+
+    current=[]
+    for value in array:
+        if isinstance(value,type(np.array([]))):
+            result=numpy_to_list(value)
+            current.append(result)
+        else:
+            
+            current.append(int(value))
+    return current
+
+
+
+@st.cache_resource
+def llm_text_response():
+    llm = Ollama(model="llama3:latest",num_ctx=1000)
+    return llm.stream
+
+def model_single_out(prompt):
+    pipe=load_model()
+    image = pipe(prompt).images[0]
+    return image
+
+def model_out_put(init_image,mask_image,prompt,negative_prompt):
+    pipeline_ = load_model()
+    image = pipeline_(prompt=prompt, negative_prompt=negative_prompt, image=init_image, mask_image=mask_image).images[0]
+    return image
+
+@st.cache_resource
+def multimodel():
+    pipeline_ = pipeline("text-classification", model = "model_collection\model_4")
+    return pipeline_
+   
+def multimodel_output(prompt):
+    pipeline_ = multimodel()
+    image = pipeline_(prompt)
+    return image[0]['label']
+
+def d4_to_3d(image):
+    formatted_array=[]
+    for j in image:
+        neste_list=[]
+        for k in j:
+            if any([True if i>0 else False for i in k]):
+                neste_list.append(True)
+            else:
+                neste_list.append(False)
+        formatted_array.append(neste_list)
+    print(np.shape(formatted_array))
+    return np.array(formatted_array)
+    
+st.set_page_config(layout="wide")
+
+
+img_selection=None
+# Specify canvas parameters in application
+drawing_mode = st.sidebar.selectbox(
+    "Drawing tool:", ("freedraw","point", "line", "rect", "circle", "transform")
+)
+
+
+dictionary=st.session_state
+if "every_prompt_with_val" not in dictionary:
+    dictionary['every_prompt_with_val']=[]
+if "current_image" not in dictionary:
+    dictionary['current_image']=[]
+if "prompt_collection"  not in dictionary:
+    dictionary['prompt_collection']=[]
+if "user" not in dictionary:
+    dictionary['user']=None
+if "current_session" not in dictionary:
+    dictionary['current_session']=None
+
+stroke_width = st.sidebar.slider("Stroke width: ", 1, 25, 20)
+if drawing_mode == 'point':
+        point_display_radius = st.sidebar.slider("Point display radius: ", 1, 25, 3)
+stroke_color = '#000000'
+bg_color = "#eee"
+
+
+column1,column2=st.columns([0.7,0.35])
+
+with open("DataBase\datetimeRecords.json","r") as read:
+    dateTimeRecord=json.load(read)
+with column2:
+    st.header("HISTORY")
+    tab1,tab2,tab3,tab4=st.tabs(["CHAT HISTORY","IMAGES","PROMPT IMPROVEMENT","LOGIN"])
+    with tab1:
+
+        
+ 
+        if not len(dictionary['every_prompt_with_val']):
+            st.header("I will store all the chat for the current session")
+            with open("lotte_animation_saver\\animation_4.json") as read:
+                url_json=json.load(read)
+            st_lottie(url_json,height = 400)
+        else:
+
+            with st.container(height=600):
+
+
+                for index,prompts_ in enumerate(dictionary['every_prompt_with_val'][::-1]):
+                    if prompts_[-1]=="@working":
+                        if index==0:
+                            st.write(prompts_[0].upper())
+                            data_need=st.write_stream(llm_text_response()(prompts_[0]))
+                            dictionary['every_prompt_with_val'][-1]=(prompts_[0],str(data_need))
+                            
+                    elif isinstance(prompts_[-1],str):
+                        if index==0:
+                            st.text_area(label=prompts_[0].upper(),value=prompts_[-1],height=500)
+                        else:
+                            st.text_area(label=prompts_[0].upper(),value=prompts_[-1])
+
+                    else:
+                        st.write(prompts_[0].upper())
+                        with st.container(height=400):
+                            format1,format2=st.columns([0.2,0.8])
+                            with format1:
+                                new_img=Image.open("ALL_image_formation\image_gen.png")
+                                st.write("<br>",unsafe_allow_html=True)
+                                size = min(new_img.size)
+                                mask = Image.new('L', (size, size), 0)
+                                draw = ImageDraw.Draw(mask)
+                                draw.ellipse((0, 0, size, size), fill=255)
+
+                                image = new_img.crop((0, 0, size, size))
+                                image.putalpha(mask)
+                                st.image(image)                    
+                            with format2:
+
+                                st.write("<br>",unsafe_allow_html=True)
+                                size = min(prompts_[-1].size)
+                                mask = Image.new('L', (size, size), 0)
+                                draw = ImageDraw.Draw(mask)
+                                draw.ellipse((0, 0, size, size), fill=255)
+
+                                # Crop the image to a square and apply the mask
+                                image = prompts_[-1].crop((0, 0, size, size))
+                                image.putalpha(mask)
+                                st.image(image)
+ 
+    with tab2:
+        
+        if "current_image" in dictionary and len(dictionary['current_image']):
+            with st.container(height=600):
+                dictinory_length=len(dictionary['current_image'])
+                
+                img_selection = image_select(
+                    label="",
+                    images=dictionary['current_image'] if len(dictionary['current_image'])!=0 else None,
+                )
+                if img_selection in dictionary['current_image']:
+                    dictionary['current_image'].remove(img_selection)
+                    dictionary['current_image'].insert(0,img_selection)
+                    # st.rerun()
+
+                img_selection.save("image.png")
+                with open("image.png", "rb") as file:
+                    downl=st.download_button(label="DOWNLOAD",data=file,file_name="image.png",mime="image/png")
+                os.remove("image.png")
+        else:
+
+            st.header("This section will store the updated images")
+            with open("lotte_animation_saver\\animation_1.json") as read:
+                url_json=json.load(read)
+            st_lottie(url_json,height = 400)
+    with tab3:
+        if len(dictionary['prompt_collection'])!=0:
+            with st.container(height=600):
+                prompt_selection=st.selectbox(label="Select the prompt for improvment",options=["Mention below are prompt history"]+dictionary["prompt_collection"],index=0)
+
+                if prompt_selection!="Mention below are prompt history":
+
+                    generated_prompt=prompt_improvment(prompt_selection)
+                    dictionary['generated_image_prompt'].append(generated_prompt)
+                    st.write_stream(generated_prompt)
+
+        else:
+
+            st.header("This section will provide prompt improvement section")
+            with open("lotte_animation_saver\\animation_3.json") as read:
+                url_json=json.load(read)
+            st_lottie(url_json,height = 400)
+        with tab4:
+            
+            # with st.container(height=600):
+
+            if not dictionary['user']   : 
+                with st.form("my_form"):
+                    # st.header("Please login for save your data")
+                    with open("lotte_animation_saver\\animation_5.json") as read:
+                        url_json=json.load(read)
+                    st_lottie(url_json,height = 200)                
+                    user_id = st.text_input("user login")
+                    password = st.text_input("password",type="password")
+                    submitted_login = st.form_submit_button("Submit")
+                    # Every form must have a submit button.
+
+                    if submitted_login:
+                        with open("DataBase\login.json","r") as read:
+                            login_base=json.load(read)
+                        if user_id in login_base and login_base[user_id]==password:
+                            dictionary['user']=user_id
+                            st.rerun()
+                        else:
+                            st.error("userid or password incorrect")
+
+                        st.write("working")
+                    modal = Modal(
+                        "Sign up", 
+                        key="demo-modal",
+                        
+                        padding=10,    # default value
+                        max_width=600  # default value
+                    )
+                open_modal = st.button("sign up")
+                if open_modal:
+                    modal.open()
+
+                if modal.is_open():
+                    with modal.container():
+
+                        with st.form("my_form1"):
+                            sign_up_column_left,sign_up_column_right=st.columns(2)
+                            with sign_up_column_left:
+                                with open("lotte_animation_saver\\animation_6.json") as read:
+                                    url_json=json.load(read)
+                                st_lottie(url_json,height = 200) 
+    
+                            with sign_up_column_right:
+                                user_id = st.text_input("user login")
+                                password = st.text_input("password",type="password")
+                                submitted_signup = st.form_submit_button("Submit")
+
+                            if submitted_signup:
+                                with open("DataBase\login.json","r") as read:
+                                    login_base=json.load(read)
+                                if not login_base:
+                                    login_base={}
+                                if user_id not in login_base:
+                                    login_base[user_id]=password
+                                    with open("DataBase\login.json","w") as write:
+                                        json.dump(login_base,write,indent=2)  
+                                    st.success("you are a part now")  
+                                    dictionary['user']=user_id
+                                    modal.close()                       
+                                else:
+                                    st.error("user id already exists")
+            else:
+                st.header("REPORTED ISSUES")
+                with st.container(height=370):
+
+                    with open("DataBase\datetimeRecords.json") as feedback:
+                        temp_issue=json.load(feedback)
+
+                    arranged_feedback=reversed(temp_issue['database'])
+                    
+                    for report in arranged_feedback:
+                        user_columns,user_feedback=st.columns([0.3,0.8])
+
+                        with user_columns:
+                            st.write(report[-1])
+                        with user_feedback:
+                            st.write(report[1])
+                     
+                feedback=st.text_area("Feedback Report and Improvement",placeholder="")
+                summit=st.button("submit")
+                if summit:
+                    with open("DataBase\datetimeRecords.json","r") as feedback_sumit:
+                        temp_issue_submit=json.load(feedback_sumit)        
+                    if  "database" not in temp_issue_submit:
+                        temp_issue_submit["database"]=[]
+                    temp_issue_submit["database"].append((str(datetime.now()),feedback,dictionary['user'])) 
+                    with open("DataBase\datetimeRecords.json","w") as feedback_sumit:
+                        json.dump(temp_issue_submit,feedback_sumit)                    
+                            
+
+
+                    # st.rerun()
+                
+                
+                    
+
+
+
+bg_image = st.sidebar.file_uploader("PLEASE UPLOAD IMAGE FOR EDITING:", type=["png", "jpg"])
+bg_doc = st.sidebar.file_uploader("PLEASE UPLOAD DOC FOR PPT/PDF/STORY:", type=["pdf","xlsx"])
+
+
+if "bg_image" not in dictionary:
+    dictionary["bg_image"]=None
+
+if img_selection  and dictionary['bg_image']==bg_image:
+    gen_image=dictionary['current_image'][0]
+else:
+    if bg_image:
+        gen_image=Image.open(bg_image) 
+    else:
+        gen_image=None
+
+
+
+
+
+
+with column1:
+# Create a canvas component
+    changes,implementation,current=st.columns([0.3,0.6,0.3])
+
+    with implementation:
+                st.write("<br>"*5,unsafe_allow_html=True)
+                canvas_result = st_canvas(
+                    fill_color="rgba(255, 165, 0, 0.3)",  # Fixed fill color with some opacity
+                    stroke_width=stroke_width,
+                    stroke_color=stroke_color,
+                    background_color=bg_color,
+                    background_image=gen_image if gen_image else Image.open("ALL_image_formation\image_gen.png"),
+                    update_streamlit=True,
+                    height=500,
+                    width=500,
+                    drawing_mode=drawing_mode,
+                    point_display_radius=point_display_radius if drawing_mode == 'point' else 0,
+                    key="canvas",
+                )
+
+
+
+
+
+with column1:
+    # prompt=st.text_area("Please provide the prompt")
+    prompt=st.chat_input("Please provide the prompt")
+    
+    negative_prompt="the black masked area"
+
+    # run=st.button("run_experiment")
+
+
+
+if canvas_result.image_data is not None:
+    if prompt:
+
+        text_or_image=multimodel_output(prompt)
+        
+        if text_or_image=="LABEL_0":
+        
+            if "generated_image_prompt" not in dictionary:
+                dictionary['generated_image_prompt']=[]
+            if prompt not in dictionary['prompt_collection'] and prompt not in dictionary['generated_image_prompt']:
+                dictionary['prompt_collection']=[prompt]+dictionary['prompt_collection']
+            new_size=np.array(canvas_result.image_data).shape[:2]
+            new_size=(new_size[-1],new_size[0])
+            if bg_image!=dictionary["bg_image"] :
+                dictionary["bg_image"]=bg_image
+                if bg_image!=None:
+                    imf=Image.open(bg_image).resize(new_size)
+                else:
+                    with open("lotte_animation_saver/animation_4.json") as read:
+                        url_json=json.load(read)        
+                    st_lottie(url_json) 
+                    imf=Image.open("ALL_image_formation\home_screen.jpg").resize(new_size)
+            else:
+                if len(dictionary['current_image'])!=0:
+                    imf=dictionary['current_image'][0]
+                else:
+                    with open("lotte_animation_saver/animation_4.json") as read:
+                        url_json=json.load(read)        
+                    st_lottie(url_json) 
+                    imf=Image.open("ALL_image_formation\home_screen.jpg")
+
+            negative_image =d4_to_3d(np.array(canvas_result.image_data))
+            if np.sum(negative_image)==0:
+                negative_image=Image.fromarray(np.where(negative_image == False, True, negative_image))
+            else:
+                negative_image=Image.fromarray(negative_image)
+            
+            modifiedValue=model_out_put(imf,negative_image,prompt,negative_prompt)
+            modifiedValue.save("ALL_image_formation/current_session_image.png")
+            dictionary['current_image']=[modifiedValue]+dictionary['current_image']
+            dictionary['every_prompt_with_val'].append((prompt,modifiedValue))
+            st.rerun()
+        else:
+            st.write("nothing importent")
+            modifiedValue="@working"
+            dictionary['every_prompt_with_val'].append((prompt,modifiedValue))
+            st.rerun()
+            # st.image(modifiedValue,width=300)
+    
+        
+        
+if canvas_result.json_data is not None:
+    objects = pd.json_normalize(canvas_result.json_data["objects"]) # need to convert obj to str because PyArrow
+    for col in objects.select_dtypes(include=['object']).columns:
+        objects[col] = objects[col].astype("str")
+

dataset_formation.ipynb

@@ -0,0 +1,524 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd\n",
+    "import json\n",
+    "import random"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "<>:1: SyntaxWarning: invalid escape sequence '\\c'\n",
+      "<>:3: SyntaxWarning: invalid escape sequence '\\d'\n",
+      "<>:1: SyntaxWarning: invalid escape sequence '\\c'\n",
+      "<>:3: SyntaxWarning: invalid escape sequence '\\d'\n",
+      "C:\\Users\\rajst\\AppData\\Local\\Temp\\ipykernel_11856\\1444736939.py:1: SyntaxWarning: invalid escape sequence '\\c'\n",
+      "  image_data=pd.read_csv(\"data_set_formation\\custom_prompts_df.csv\")\n",
+      "C:\\Users\\rajst\\AppData\\Local\\Temp\\ipykernel_11856\\1444736939.py:3: SyntaxWarning: invalid escape sequence '\\d'\n",
+      "  with open(\"data_set_formation\\data.json\") as read:\n"
+     ]
+    }
+   ],
+   "source": [
+    "image_data=pd.read_csv(\"data_set_formation\\custom_prompts_df.csv\")\n",
+    "\n",
+    "with open(\"data_set_formation\\data.json\") as read:\n",
+    "    text_data=json.load(read)\n",
+    "# prompt_data="
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>prompt</th>\n",
+       "      <th>image_file</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>painting of King Henry VIII carrying an umbrella</td>\n",
+       "      <td>images/0/custom_0_0.png</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>Fox Mulder and a chinchilla walking down a roa...</td>\n",
+       "      <td>images/0/custom_1_0.png</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>photo of a gas burner by a soft pretzel</td>\n",
+       "      <td>images/0/custom_2_0.png</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>photo of Shyster standing street lights on at ...</td>\n",
+       "      <td>images/0/custom_3_0.png</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>cute young man eating a plant over a fence in ...</td>\n",
+       "      <td>images/0/custom_5_0.png</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>99995</th>\n",
+       "      <td>photo of a natural kite at Westminster Abbey</td>\n",
+       "      <td>images/102/custom_102419_0.png</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>99996</th>\n",
+       "      <td>smooth rum with a clock in the style of a digi...</td>\n",
+       "      <td>images/102/custom_102420_0.png</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>99997</th>\n",
+       "      <td>a lovable elephant by the Gamla Stan, Stockholm</td>\n",
+       "      <td>images/102/custom_102421_0.png</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>99998</th>\n",
+       "      <td>photo of Courtney Love with a hot dog</td>\n",
+       "      <td>images/102/custom_102422_0.png</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>99999</th>\n",
+       "      <td>Maniac jumping on a skateboard near a fence</td>\n",
+       "      <td>images/102/custom_102423_0.png</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>100000 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                  prompt  \\\n",
+       "0       painting of King Henry VIII carrying an umbrella   \n",
+       "1      Fox Mulder and a chinchilla walking down a roa...   \n",
+       "2                photo of a gas burner by a soft pretzel   \n",
+       "3      photo of Shyster standing street lights on at ...   \n",
+       "4      cute young man eating a plant over a fence in ...   \n",
+       "...                                                  ...   \n",
+       "99995       photo of a natural kite at Westminster Abbey   \n",
+       "99996  smooth rum with a clock in the style of a digi...   \n",
+       "99997    a lovable elephant by the Gamla Stan, Stockholm   \n",
+       "99998              photo of Courtney Love with a hot dog   \n",
+       "99999        Maniac jumping on a skateboard near a fence   \n",
+       "\n",
+       "                           image_file  \n",
+       "0             images/0/custom_0_0.png  \n",
+       "1             images/0/custom_1_0.png  \n",
+       "2             images/0/custom_2_0.png  \n",
+       "3             images/0/custom_3_0.png  \n",
+       "4             images/0/custom_5_0.png  \n",
+       "...                               ...  \n",
+       "99995  images/102/custom_102419_0.png  \n",
+       "99996  images/102/custom_102420_0.png  \n",
+       "99997  images/102/custom_102421_0.png  \n",
+       "99998  images/102/custom_102422_0.png  \n",
+       "99999  images/102/custom_102423_0.png  \n",
+       "\n",
+       "[100000 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "image_data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "data_dict={\"prompt\":[],\"label\":[]}"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "queries = [\n",
+    "    # General Descriptions\n",
+    "    \"Generate a beautiful sunset over the ocean.\",\n",
+    "    \"Create a futuristic cityscape at night.\",\n",
+    "    \"Show a cozy cabin in the middle of a snowy forest.\",\n",
+    "    \"Draw a tropical beach with palm trees and clear blue water.\",\n",
+    "    \"Design a medieval castle on a hilltop.\",\n",
+    "    \n",
+    "    # Character-Focused Queries\n",
+    "    \"Generate a young woman with long red hair in a fantasy setting.\",\n",
+    "    \"Create a warrior in futuristic armor holding a glowing sword.\",\n",
+    "    \"Draw a friendly robot helping people in a park.\",\n",
+    "    \"Design a wise old wizard with a long beard and staff.\",\n",
+    "    \"Illustrate a child playing with a puppy in a garden.\",\n",
+    "    \n",
+    "    # Animal and Nature Queries\n",
+    "    \"Show a majestic tiger in a dense jungle.\",\n",
+    "    \"Create a flock of birds flying over a mountain range.\",\n",
+    "    \"Draw a koi fish pond with colorful fish.\",\n",
+    "    \"Generate a close-up of a butterfly on a flower.\",\n",
+    "    \"Illustrate a desert landscape with cacti and a setting sun.\",\n",
+    "    \n",
+    "    # Architectural and Object Queries\n",
+    "    \"Design a futuristic spaceship hovering above Earth.\",\n",
+    "    \"Create a vintage car driving on a country road.\",\n",
+    "    \"Draw a small café on a busy European street.\",\n",
+    "    \"Generate a treehouse in the middle of a forest.\",\n",
+    "    \"Show a steampunk-style clock tower.\",\n",
+    "    \n",
+    "    # Abstract or Conceptual Queries\n",
+    "    \"Create an image representing the concept of time.\",\n",
+    "    \"Design a surreal landscape with floating islands.\",\n",
+    "    \"Generate an artwork of colors blending like a rainbow.\",\n",
+    "    \"Illustrate the feeling of calmness in visual form.\",\n",
+    "    \"Show a dreamlike city made of crystal.\",\n",
+    "    \n",
+    "    # Cultural or Historical Themes\n",
+    "    \"Illustrate an ancient Egyptian pyramid under the stars.\",\n",
+    "    \"Show a samurai in traditional armor standing in a bamboo forest.\",\n",
+    "    \"Draw a Viking ship sailing through a storm.\",\n",
+    "    \"Create an Indian temple with intricate carvings.\",\n",
+    "    \"Generate a Renaissance-style painting of a feast.\",\n",
+    "    \n",
+    "    # Event or Scene Queries\n",
+    "    \"Show a birthday party with balloons and a cake.\",\n",
+    "    \"Create an image of people camping under the stars.\",\n",
+    "    \"Draw a bustling market in a small village.\",\n",
+    "    \"Illustrate a concert with a crowd and colorful lights.\",\n",
+    "    \"Generate an image of a wedding ceremony by the beach.\",\n",
+    "    \n",
+    "    # Seasonal and Holiday Themes\n",
+    "    \"Show a Christmas scene with a decorated tree and snow.\",\n",
+    "    \"Generate a spooky Halloween setting with pumpkins and ghosts.\",\n",
+    "    \"Create a spring meadow full of flowers and butterflies.\",\n",
+    "    \"Draw an autumn forest with falling leaves.\",\n",
+    "    \"Illustrate a New Year celebration with fireworks.\"\n",
+    "]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "for i in queries:\n",
+    "    data_dict['prompt'].append(i.lower())\n",
+    "    data_dict['label'].append(\"image\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "counter=0\n",
+    "detail_list=[\"painting\",\"image\",\"photo\",\"frame\",\"picture\",\"potrait\",\"pic\",\"snapshot\"]\n",
+    "for i in image_data['prompt']:\n",
+    "    if any([paint_key in i for paint_key in [\"painting\",\"image\",\"photo\",\"frame\",\"picture\",\"potrait\",\"pic\",\"snapshot\"]]):\n",
+    "        data_dict['prompt'].append(i.lower().replace(random.choice(detail_list),\"image\"))\n",
+    "        data_dict['label'].append(\"image\")\n",
+    "        counter+=1\n",
+    "    if counter==20000:\n",
+    "        break\n",
+    "counter=0\n",
+    "for j in text_data[:20000]:\n",
+    "    data_dict['prompt'].append(j['note'].lower())\n",
+    "    data_dict['label'].append(\"text\")\n",
+    "    counter+=1\n",
+    "    if counter==15000:\n",
+    "        break"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "counter=0\n",
+    "for z in text_data[15000:]:\n",
+    "    if any([paint_key in z['note'] for paint_key in [\"painting\",\"image\",\"photo\",\"frame\",\"picture\",\"potrait\",\"pic\",\"snapshot\"]]):\n",
+    "        data_dict['prompt'].append(z['note'].lower())\n",
+    "        data_dict['label'].append(\"text\")\n",
+    "        counter+=1\n",
+    "    if counter==5000:\n",
+    "        break"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>prompt</th>\n",
+       "      <th>label</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>generate a beautiful sunset over the ocean.</td>\n",
+       "      <td>image</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>create a futuristic cityscape at night.</td>\n",
+       "      <td>image</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>show a cozy cabin in the middle of a snowy for...</td>\n",
+       "      <td>image</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>draw a tropical beach with palm trees and clea...</td>\n",
+       "      <td>image</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>design a medieval castle on a hilltop.</td>\n",
+       "      <td>image</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>40035</th>\n",
+       "      <td>i was watching a documentary and it spoke of s...</td>\n",
+       "      <td>text</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>40036</th>\n",
+       "      <td>should i buy a dslr or a new phone for photogr...</td>\n",
+       "      <td>text</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>40037</th>\n",
+       "      <td>okay, i see. so it depends on how serious i am...</td>\n",
+       "      <td>text</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>40038</th>\n",
+       "      <td>it is just to take photos of my family</td>\n",
+       "      <td>text</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>40039</th>\n",
+       "      <td>is there any topical treatment i can apply to ...</td>\n",
+       "      <td>text</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>40040 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                  prompt  label\n",
+       "0            generate a beautiful sunset over the ocean.  image\n",
+       "1                create a futuristic cityscape at night.  image\n",
+       "2      show a cozy cabin in the middle of a snowy for...  image\n",
+       "3      draw a tropical beach with palm trees and clea...  image\n",
+       "4                 design a medieval castle on a hilltop.  image\n",
+       "...                                                  ...    ...\n",
+       "40035  i was watching a documentary and it spoke of s...   text\n",
+       "40036  should i buy a dslr or a new phone for photogr...   text\n",
+       "40037  okay, i see. so it depends on how serious i am...   text\n",
+       "40038             it is just to take photos of my family   text\n",
+       "40039  is there any topical treatment i can apply to ...   text\n",
+       "\n",
+       "[40040 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 13,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "pd.DataFrame(data_dict)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "folder_path = 'formatted_data/'\n",
+    "\n",
+    "# Get the list of all files in the folder\n",
+    "file_names = os.listdir(folder_path)\n",
+    "max_file_name=max([int(i.split(\"_\")[-1][:-4]) for i in file_names])"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Confussing prompts"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# from langchain_community.llms import Ollama\n",
+    "# llm = Ollama(model=\"llava:34b \",num_ctx=10000)\n",
+    "# enhancement=\"I need to train a model to distinguish between text and images. Please create a list of challenging prompts where the model needs to decide whether to generate text or identify an image.\"\n",
+    "# prompt = enhancement\n",
+    "# # result = llm.invoke(prompt)\n",
+    "# value=llm.invoke(prompt)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# print(str(value))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pd.DataFrame(data_dict).to_csv(\"formatted_data/data_\"+str(max_file_name+1)+\".csv\",index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "env",
+   "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": "3.12.4"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}