Upload 9 files

Textual Dataset Generation/.idea/.gitignore

@@ -0,0 +1,8 @@
+# Default ignored files
+/shelf/
+/workspace.xml
+# Editor-based HTTP Client requests
+/httpRequests/
+# Datasource local storage ignored files
+/dataSources/
+/dataSources.local.xml

Textual Dataset Generation/.idea/Textual Dataset Generation.iml

@@ -0,0 +1,9 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<module type="JAVA_MODULE" version="4">
+  <component name="NewModuleRootManager" inherit-compiler-output="true">
+    <exclude-output />
+    <content url="file://$MODULE_DIR$" />
+    <orderEntry type="inheritedJdk" />
+    <orderEntry type="sourceFolder" forTests="false" />
+  </component>
+</module>

Textual Dataset Generation/.idea/misc.xml

@@ -0,0 +1,6 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectRootManager" version="2" languageLevel="JDK_11" default="true" project-jdk-name="11" project-jdk-type="JavaSDK">
+    <output url="file://$PROJECT_DIR$/out" />
+  </component>
+</project>

Textual Dataset Generation/.idea/modules.xml

@@ -0,0 +1,8 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectModuleManager">
+    <modules>
+      <module fileurl="file://$PROJECT_DIR$/.idea/Textual Dataset Generation.iml" filepath="$PROJECT_DIR$/.idea/Textual Dataset Generation.iml" />
+    </modules>
+  </component>
+</project>

Textual Dataset Generation/.idea/workspace.xml

@@ -0,0 +1,83 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ChangeListManager">
+    <list default="true" id="788fdcfa-b47d-4627-847f-ec90b9a0d9dd" name="Changes" comment="" />
+    <option name="SHOW_DIALOG" value="false" />
+    <option name="HIGHLIGHT_CONFLICTS" value="true" />
+    <option name="HIGHLIGHT_NON_ACTIVE_CHANGELIST" value="false" />
+    <option name="LAST_RESOLUTION" value="IGNORE" />
+  </component>
+  <component name="ProjectColorInfo"><![CDATA[{
+  "associatedIndex": 5
+}]]></component>
+  <component name="ProjectId" id="2hyRbehAgzDFNKMNHP7EtzBH3c2" />
+  <component name="ProjectViewState">
+    <option name="hideEmptyMiddlePackages" value="true" />
+    <option name="showLibraryContents" value="true" />
+  </component>
+  <component name="PropertiesComponent"><![CDATA[{
+  "keyToString": {
+    "Python.web scrap.executor": "Run",
+    "RunOnceActivity.OpenProjectViewOnStart": "true",
+    "RunOnceActivity.ShowReadmeOnStart": "true",
+    "kotlin-language-version-configured": "true",
+    "last_opened_file_path": "C:/Users/Amir/Desktop/Textual Dataset Generation",
+    "nodejs_package_manager_path": "npm",
+    "project.structure.last.edited": "SDKs",
+    "project.structure.proportion": "0.15",
+    "project.structure.side.proportion": "0.22758621",
+    "vue.rearranger.settings.migration": "true"
+  }
+}]]></component>
+  <component name="RunManager">
+    <configuration name="web scrap" type="PythonConfigurationType" factoryName="Python" nameIsGenerated="true">
+      <module name="Textual Dataset Generation" />
+      <option name="ENV_FILES" value="" />
+      <option name="INTERPRETER_OPTIONS" value="" />
+      <option name="PARENT_ENVS" value="true" />
+      <envs>
+        <env name="PYTHONUNBUFFERED" value="1" />
+      </envs>
+      <option name="SDK_HOME" value="" />
+      <option name="SDK_NAME" value="Python 3.11 (base)" />
+      <option name="WORKING_DIRECTORY" value="$PROJECT_DIR$" />
+      <option name="IS_MODULE_SDK" value="false" />
+      <option name="ADD_CONTENT_ROOTS" value="true" />
+      <option name="ADD_SOURCE_ROOTS" value="true" />
+      <EXTENSION ID="PythonCoverageRunConfigurationExtension" runner="coverage.py" />
+      <option name="SCRIPT_NAME" value="$PROJECT_DIR$/web scrap.py" />
+      <option name="PARAMETERS" value="" />
+      <option name="SHOW_COMMAND_LINE" value="false" />
+      <option name="EMULATE_TERMINAL" value="false" />
+      <option name="MODULE_MODE" value="false" />
+      <option name="REDIRECT_INPUT" value="false" />
+      <option name="INPUT_FILE" value="" />
+      <method v="2" />
+    </configuration>
+  </component>
+  <component name="SharedIndexes">
+    <attachedChunks>
+      <set>
+        <option value="jdk-11.0.21-corretto-11.0.21-3183f394aec4-d55de845" />
+      </set>
+    </attachedChunks>
+  </component>
+  <component name="SpellCheckerSettings" RuntimeDictionaries="0" Folders="0" CustomDictionaries="0" DefaultDictionary="application-level" UseSingleDictionary="true" transferred="true" />
+  <component name="TaskManager">
+    <task active="true" id="Default" summary="Default task">
+      <changelist id="788fdcfa-b47d-4627-847f-ec90b9a0d9dd" name="Changes" comment="" />
+      <created>1718564403600</created>
+      <option name="number" value="Default" />
+      <option name="presentableId" value="Default" />
+      <updated>1718564403600</updated>
+      <workItem from="1718564404645" duration="101000" />
+    </task>
+    <servers />
+  </component>
+  <component name="TypeScriptGeneratedFilesManager">
+    <option name="version" value="3" />
+  </component>
+  <component name="com.intellij.coverage.CoverageDataManagerImpl">
+    <SUITE FILE_PATH="coverage/Textual_Dataset_Generation$web_scrap.coverage" NAME="web scrap Coverage Results" MODIFIED="1718564477196" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$" />
+  </component>
+</project>

Textual Dataset Generation/Dataset/input_dataset.csv

@@ -0,0 +1,161 @@
+category,caption
+shoes,a person sitting on top of a wooden bench
+shoes,unpaired red Nike sneaker
+shoes,brown Nike sneaker on yellow textile
+shoes,unpaired maroon plimsoll on top of yellow textile
+shoes,a person with their feet up in the air
+shoes,white and blue nike air force 1 high
+shoes,pair of white-and-orange athletic shoes on white box
+shoes,green and black nike athletic shoe
+shoes,"Various colorful sports shoes laid on sand beach background, studio shot, flat lay."
+shoes,unpaired OFF WHITE X Nike Air Force 1 low-top sneaker
+shoes,white and red nike athletic shoe
+shoes,person wearing white Nike running shoes standing on black concrete path
+shoes,unpaired gray Nike running shoe
+shoes,pair of blue-and-pink floral almond-toe pumps
+shoes,pair of gray running shoes
+shoes,Close up of a young skater girl's feet and skateboard
+shoes,white black and red nike air max 90
+shoes,a pair of white and brown shoes on a table
+shoes,unpaired red Air Jordan 12
+sneakers,a man sitting on a brick wall next to a skateboard
+sneakers,a person wearing a hat
+sneakers,black white and red nike high top sneaker
+sneakers,two people sitting on the ground with their legs crossed
+sneakers,pair of black-white-and-red Air Jordan 1 shoes
+sneakers,Nike shoe lot
+sneakers,white nike air force 1 low
+sneakers,person wearing white nike sneakers
+sneakers,a man holding a pair of purple shoes
+sneakers,close-up photography of person wears brown-and-white Nike Air Max
+sneakers,person holding white and red heart print box
+sneakers,pair of Carhartt x Nike Air Force 1 shoes
+sneakers,Legs and sneakers of teenage boys and girls standing on the sidewalk
+sneakers,man sitting on the ledge of a building wearing Air Jordan 1 low-top shoes
+sneakers,blue white and red neon light
+sneakers,person wearing black pants and blue and white nike sneakers
+heels,a person standing on top of a wooden stool
+heels,woman in pink patent leather stilettos
+heels,woman in black leather heeled shoes
+heels,a pair of shoes sitting on top of a couch
+heels,red and white plastic frame on white table
+heels,pair of women's brown pointed-toe pumps on board
+heels,woman wearing brown leather heeled sandals walking on staircase
+heels,a pair of woman's legs wearing black gloves and red high heels
+heels,a pair of purple high heeled shoes on a woman's leg
+heels,person in black pants and red shoes
+heels,womens white and silver peep toe pumps
+heels,a woman wearing red high heels standing on a wooden floor
+heels,black leather boot on white surface
+heels,gray leather peep toe sandals
+heels,brown leather boots on white table
+heels,a woman's legs wearing red high heels
+heels,black leather peep toe heeled shoes
+heels,women's seven assorted-color footwear on surface
+heels,womens brown leather peep toe heeled shoes
+watches,a man is holding onto a railing in a room
+watches,blue and silver analog watch at 10 00
+watches,gold and white analog watch
+watches,round silver-colored analog watch with black strap
+watches,Close up of man putting watch on his hand.
+watches,brown and white analog watch at 10 10
+watches,round silver-colored watch on rack during sunset
+watches,silver and black chronograph watch
+watches,a person holding a cup of tea in their hand
+watches,round gray analog watch with brown band
+watches,watch at 10:34
+watches,silver and white round analog watch
+watches,Man in black suit wear new watches. Luxury style
+watches,person wearing silver link bracelet round analog watch
+watches,round silver-colored Nomos watch at 10:10
+watches,silver link bracelet round chronograph watch
+watches,a woman sitting at a table with a cell phone in her hand
+watches,black and silver round analog watch
+watches,watch at 8:45
+watches,blue and gold analog watch
+pants,three pairs of jeans are lined up on a white surface
+pants,person in pink pants and white shoes
+pants,blue denim jeans on white textile
+pants,person wears blue jeans
+pants,a pair of black leather boots sitting on top of a pile of blue jeans
+pants,man wearing brown fitted jeans and sneakers standing on road at daytime
+pants,three assorted-color denim bottoms
+pants,blue denim jeans on brown clothes hanger
+pants,a man walking across a bridge carrying a suitcase
+pants,man in white t-shirt and black pants standing on white floor
+pants,woman in orange pants
+pants,selective focus photography of hanged denim jeans
+pants,a person sitting on a bench with their legs crossed
+pants,woman in blue denim jeans and white sneakers
+pants,blue denim jeans on black surface
+pants,a pair of hands holding a pair of jeans
+pants,closeup photo of person hiding his right hand in his pocket
+pants,woman in white tank top and blue denim jeans standing on beach during daytime
+pants,woman in gray tank top and gray pants
+clothing,a piece of cloth sitting on top of a table
+clothing,"gray cardigan, blue jeans, and pair of brown chunky heeled shoes"
+clothing,green clothes hanger
+clothing,closeup of hanged shirts on rack
+clothing,a woman holding a pair of jeans
+clothing,women's white long sleeve shirt
+clothing,pair of white low-top sneakers
+clothing,hanged jeans lot
+clothing,a close up of a pair of blue jeans
+clothing,gray dress shirt hang on brown wooden rack in front of window with white curtain
+clothing,white and red nike air force 1 high
+clothing,woman standing selecting clothes
+clothing,a stack of folded clothes on a table next to a lamp
+clothing,woman waering black blouse and hat
+clothing,woman in red long sleeve dress
+clothing,assorted color folded shirts on wooden panel
+clothing,a pile of folded clothes sitting on top of a brown leather chair
+clothing,selective focus photography of hanged three gray tee shirts
+clothing,white and blue cat-printed crew-neck T-shirt
+clothing,assorted-color shirt lot hang on rack
+dress,a woman sitting on a bed in a room
+dress,woman in red sleeveless dress standing on gray concrete floor during daytime
+dress,green sleeveless dress hanged on white wall
+dress,woman in seashore
+dress,a woman standing on a beach holding a banana leaf
+dress,woman walking on seaside while holding woven bag
+dress,woman in yellow and white floral dress
+dress,woman wearing black turtleneck long-sleeved dress
+dress,a woman kneeling on a white background posing for a picture
+dress,woman walking down stair under clear blue sky during daytime
+dress,black leather spaghetti strap dress
+dress,woman standing in front of white wall
+dress,woman in beige floral sleeveless dress
+dress,woman wearing pink dress while standing near black metal rail at daytime
+dress,a rack of clothes hanging on a wall
+dress,woman in brown long sleeve dress standing beside gray wall
+dress,woman in green long sleeve dress standing on brown field during daytime
+dress,selective focus photo of smiling woman wearing black dress standing on concrete pavement
+shirt,a close up of an umbrella with a metal handle
+shirt,white crew neck t-shirt
+shirt,man in white dress shirt wearing black sunglasses
+shirt,gray button up long sleeve shirt
+shirt,"A young businessman standing in corridor outside office, looking at camera."
+shirt,black crew neck t-shirt
+shirt,blue and white checkered dress shirt
+shirt,man in black and white plaid button up shirt wearing black sunglasses
+shirt,white button up shirt on clothes hanger
+shirt,white crew neck long sleeve shirt
+shirt,A fathers day greeting card concept. Flat lay. Copy space.
+shirt,man in white dress shirt with black and white polka dots necktie
+shirt,woman wearing orange crew-neck sweatshirt standing while putting right hand on her head
+shirt,a person holding a white shirt on a hanger
+shirt,man in blue dress shirt wearing black sunglasses
+shirt,man in white crew neck t-shirt standing on green grass field during daytime
+t-shirt,a young man wearing a black shirt and a chain around his neck
+t-shirt,man wearing white crew-neck t-shirts
+t-shirt,a woman sitting on the floor with her legs crossed
+t-shirt,man in white crew neck t-shirt
+t-shirt,smiling woman in black and white print t-shirt
+t-shirt,woman in black crew neck t-shirt and blue denim jeans
+t-shirt,photo of blue crew-neck tops
+t-shirt,a woman in a white shirt and a pink hat
+t-shirt,woman in black crew neck t-shirt standing near painting
+t-shirt,men's white crew-neck t-shirt
+t-shirt,man wearing white crew-neck t-shirt
+t-shirt,man in pink crew neck t-shirt wearing black sunglasses
+t-shirt,man in black crew neck t-shirt standing during daytime

Textual Dataset Generation/main.py

@@ -0,0 +1,256 @@
+import json
+import pandas as pd
+
+color_changes = [
+    ("black", "Change * color to black"),
+    ("white", "Change * color to white"),
+    ("red", "Change * color to red"),
+    ("blue", "Change * color to blue"),
+    ("orange", "Change * color to orange"),
+    ("purple", "Change * color to purple"),
+]
+
+scene_changes = [
+    ("by the beach", "put * by the beach"),
+    ("in the desert", "put * in the desert"),
+    ("* in a festival", "put it in a festival"),
+    ("on a table", "put * on a table"),
+    ("in a forest", "put * in a forest"),
+    ("on a roof", "put * on a roof"),
+    ("in the air", "make * in the air"),
+    ("in hands", "put * in someone`s hands"),
+    ("covered with water", "put * in water"),
+]
+
+season_changes = [
+    ("in snow", "add snow effects"),
+    ("in spring", "add spring effects"),
+    ("in summer", "add summer effects"),
+    ("in autumn", "add autumn effects"),
+    ("during daytime", "make the scene in daytime"),
+    ("during nighttime", "make the period in nighttime"),
+]
+
+background_changes = [
+    ("in front of a mountain", "add mountain in the background"),
+    ("in front of the pyramids", "add the pyramids in the background"),
+]
+
+PRODUCTS = [
+    "t-shirt",
+    "shirt",
+    "dress",
+    "pants",
+    "shoe",
+    "watch",
+    "vase",
+    "sneaker",
+    "headphone",
+    "bottle",
+    "perfume",
+    "vase",
+    "cup",
+    "camera",
+    "phone",
+    "mobile",
+    "bag",
+    "earpod",
+    "earbud",
+    "heel",
+]
+
+
+def find_words(lst, input):
+    found_words = []
+    for w in lst:
+        indx = input.find(w)
+        if indx != -1:
+            if indx + len(w) + 1 < len(input) and input[indx + len(w) + 1] == "s":
+                w += "s"
+
+            found_words.append(w)
+
+    return found_words
+
+
+def perform_color_changes(input, mixed_color=False):
+    if input.endswith("s"):
+        input = input[:-1]
+
+    original_input = input
+    result = []
+    found_words = find_words(PRODUCTS, input)
+
+    # remove colors to avoid repeated consecutive colors
+    for change, _ in color_changes:
+        changee = " "
+        if mixed_color == True:
+            changee = change + "-"
+
+        input = input.replace(str(" " + change + " "), changee).replace("  ", " ")
+
+    for word in found_words:
+        for change, change_text in color_changes:
+            edit = change_text.replace("*", word)
+            changed = str(change + " " + word)
+            output = input.replace(word, changed)
+
+            result.append({"caption": original_input, "edit": edit, "output": output})
+
+    return result
+
+
+def perform_scene_changes(input):
+    original_input = input
+    for change, _ in scene_changes:
+        input = input.replace(change, "")
+
+    result = []
+    found_words = find_words(PRODUCTS, input)
+    for word in found_words:
+        for change, change_text in scene_changes:
+            edit = change_text
+            edit = edit.replace("*", word)
+
+            output = input + " " + change
+
+            result.append(
+                {
+                    "caption": original_input,
+                    "edit": edit,
+                    "output": output.replace("  ", " "),
+                }
+            )
+
+    return result
+
+
+def perform_affects_changes(input):
+    original_input = input
+    for change, _ in background_changes:
+        input = input.replace(change, "")
+
+    result = []
+    found_words = find_words(PRODUCTS, input)
+
+    for _ in found_words:
+        for change, change_text in season_changes:
+            edit = change_text
+            output = input + " " + change
+
+            result.append(
+                {
+                    "caption": original_input,
+                    "edit": edit,
+                    "output": output.replace("  ", " "),
+                }
+            )
+
+        break
+    return result
+
+
+def perform_background_changes(input):
+    original_input = input
+    for change, _ in background_changes:
+        input = input.replace(change, "")
+
+    result = []
+    found_words = find_words(PRODUCTS, input)
+
+    for _ in found_words:
+        for change, change_text in background_changes:
+            edit = change_text
+            output = input + " " + change
+
+            result.append(
+                {
+                    "caption": original_input,
+                    "edit": edit,
+                    "output": output.replace("  ", " "),
+                }
+            )
+        break
+
+    return result
+
+
+def get_rows(input):  # return: list of dictionary
+    input = input.lower().replace("-", " ")
+    # remove brands
+    input = input.replace("nike ", "")
+    result = []
+
+    color_changes_result = perform_color_changes(input, False)
+    for new_row in color_changes_result:
+        result.append(new_row)
+
+    scene_changes_result = perform_scene_changes(input)
+    for new_row in scene_changes_result:
+        result.append(new_row)
+
+    affects_changes_result = perform_affects_changes(input)
+    for new_row in affects_changes_result:
+        result.append(new_row)
+
+    background_changes_result = perform_background_changes(input)
+    for new_row in background_changes_result:
+        result.append(new_row)
+
+    return result
+
+
+def statistics(strs):
+    total = 0
+    stat = {}
+    for product in PRODUCTS:
+        cntr = sum(product.lower() in s.lower() for s in strs)
+        total += cntr
+        stat[product] = cntr
+
+    stat = dict(sorted(stat.items(), key=lambda item: item[1], reverse=True))
+
+    return stat, total
+
+
+def load_dataset(path):
+
+    dataset = pd.read_csv(path)
+    return dataset
+
+
+def save_dataset(data, file_path):
+    with open(file_path, "w") as file:
+        for idx in range(len(data)):
+            json_string = json.dumps(data[idx]) + ("\n" if idx < len(data) - 1 else "")
+            file.write(json_string)
+
+
+def insert_row(df, new_data: list):
+    # Create a DataFrame with the new row data
+    new_data = pd.DataFrame(new_data, columns=df.columns)
+    # Append the new row to the existing DataFrame
+    df = df.append(new_data, ignore_index=True)
+    return df
+
+
+def main():
+    input_dataset = load_dataset("Dataset/input_dataset.csv")
+    outout_dataset = []
+
+    for val in input_dataset["caption"]:
+        newData = get_rows(val)
+        if len(newData) > 0:
+            outout_dataset += newData
+
+    stat, total = statistics([item["caption"] for item in outout_dataset])
+    stat = pd.DataFrame([stat]).transpose()
+
+    print(rf"TOTAL : {total}")
+    print(stat)
+
+    save_dataset(outout_dataset, "Dataset/textual_dataset.jsonl")
+
+
+if __name__ == "__main__":
+    main()

Textual Dataset Generation/web scrap.py

@@ -0,0 +1,66 @@
+import pandas as pd
+import requests
+from bs4 import BeautifulSoup
+import time
+
+
+def scrap(url: str):
+    try:
+        response = requests.get(url)
+        response.raise_for_status()  # Raise an error for bad status codes
+    except requests.exceptions.RequestException as e:
+        print(f"Error fetching {url}: {e}")
+        return None
+
+    # Parse the HTML content
+    soup = BeautifulSoup(response.content, "html.parser")
+
+    div_elements = soup.find_all("div", class_="WxXog")
+    result = []
+    for div in div_elements:
+        img_tag = div.find("img")
+        if img_tag:
+            alt_text = img_tag.get("alt")
+            if alt_text is not None and len(alt_text) < 100:
+                result.append(alt_text)
+
+    return result
+
+
+def main():
+    data = []
+
+    categories = [
+        "shoes",
+        "sneakers",
+        "heels",
+        "watches",
+        "pants",
+        "clothing",
+        "dress",
+        "shirt",
+        "t-shirt",
+    ]
+    seen_inputs = set()
+    base_url = "https://unsplash.com/s/photos/"
+
+    for category in categories:
+        url = f"{base_url}{category}"
+        print(f"Scraping {url}...")
+        res = scrap(url)
+        if res is not None:
+            for t in res:
+                if t not in seen_inputs:
+                    data.append({"category": category, "caption": t})
+                    seen_inputs.add(t)
+
+        time.sleep(1)  # Delay to avoid rate limiting
+
+    df = pd.DataFrame(data)
+
+    # Save to CSV
+    df.to_csv("input_dataset.csv", index=False)
+
+
+if __name__ == "__main__":
+    main()