File size: 12,125 Bytes
861698e
 
7348981
 
 
 
 
861698e
 
 
7348981
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
861698e
f234c19
7348981
 
 
 
 
7437080
7348981
 
 
 
 
 
 
 
 
 
1172ef2
861698e
 
 
7348981
 
 
861698e
 
7348981
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
861698e
7348981
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ef744fb
7348981
 
 
 
 
 
 
 
 
 
 
 
 
ef744fb
7348981
 
 
 
 
 
 
 
 
ef744fb
7348981
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ef744fb
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
import openai
import os
import gradio as gr
from enum import Enum
from dataclasses import dataclass, asdict, field
from typing import List, Optional, Union, Dict, Any
import json

api_key = os.getenv("OPEN_AI_KEY")

# Define the COCOClass enum
class COCOClass(Enum):
    person = 0
    bicycle = 1
    car = 2
    motorcycle = 3
    airplane = 4
    bus = 5
    train = 6
    truck = 7
    boat = 8
    traffic_light = 9
    fire_hydrant = 10
    stop_sign = 11
    parking_meter = 12
    bench = 13
    bird = 14
    cat = 15
    dog = 16
    horse = 17
    sheep = 18
    cow = 19
    elephant = 20
    bear = 21
    zebra = 22
    giraffe = 23
    backpack = 24
    umbrella = 25
    handbag = 26
    tie = 27
    suitcase = 28
    frisbee = 29
    skis = 30
    snowboard = 31
    sports_ball = 32
    kite = 33
    baseball_bat = 34
    baseball_glove = 35
    skateboard = 36
    surfboard = 37
    tennis_racket = 38
    bottle = 39
    wine_glass = 40
    cup = 41
    fork = 42
    knife = 43
    spoon = 44
    bowl = 45
    banana = 46
    apple = 47
    sandwich = 48
    orange = 49
    broccoli = 50
    carrot = 51
    hot_dog = 52
    pizza = 53
    donut = 54
    cake = 55
    chair = 56
    couch = 57
    potted_plant = 58
    bed = 59
    dining_table = 60
    toilet = 61
    tv = 62
    laptop = 63
    mouse = 64
    remote = 65
    keyboard = 66
    cell_phone = 67
    microwave = 68
    oven = 69
    toaster = 70
    sink = 71
    refrigerator = 72
    book = 73
    clock = 74
    vase = 75
    scissors = 76
    teddy_bear = 77
    hair_drier = 78
    toothbrush = 79

# Define data classes
@dataclass
class VehicleProps:
    brand: Optional[str] = None
    type: Optional[COCOClass] = None  # Should be a vehicle class
    plate: Optional[str] = None

@dataclass
class PersonProps:
    face_images: Optional[List[str]] = field(default_factory=list)
    age: Optional[int] = None
    race: Optional[str] = None  # Should be one of the specified races
    gender: Optional[str] = None  # Male or Female
    top_color: Optional[str] = None  # Changed from shirt_color
    bottom_color: Optional[str] = None

@dataclass
class Activity:
    prompt: Optional[str] = None
    type: Optional[str] = None  # "full_screen" or "square"

@dataclass
class Investigation:
    target: COCOClass
    images: List[str]
    activity: Optional[Activity] = None
    complex_appearance: Optional[str] = None
    props: Optional[Union[VehicleProps, PersonProps]] = None
    primary_color: Optional[str] = None
    secondary_color: Optional[str] = None

# Default system message (moved to a global variable)
DEFAULT_SYSTEM_MESSAGE = """
You are a helpful assistant that extracts structured information from text descriptions.

Your task is to parse the following text prompt and extract information to populate an Investigation JSON object as per the definitions provided.

Definitions:

Investigation:
{{
    "target": A COCO class name (from the COCOClass enum),
    "images": List of image URLs,
    "activity": {{
        "prompt": A description of an activity, e.g., "crossing the street", "crossing red light", "holding a gun",
        "type": Either "full_screen" or "square"
            - "full_screen": When the activity requires the full scene for context (e.g., "seeing a movie").
            - "square": When the activity context can be understood from a close-up image (e.g., "holding a cat").
    }},
    "complex_appearance": Description of appearance details that do not fit into other fields, e.g., "has a hat with Nike logo" or "Tattoo on left arm",
    "props": Either VehicleProps or PersonProps (only if the target is vehicle or person),
    "primary_color": Primary color mentioned in the prompt,
    "secondary_color": Secondary color mentioned in the prompt
}}

VehicleProps:
{{
    "brand": Vehicle brand, e.g., "Mercedes",
    "type": COCO class name of vehicles (e.g., "truck"),
    "plate": License plate number, e.g., "123AB"
}}

PersonProps:
{{
    "face_images": List of face image URLs,
    "age": Age as a number,
    "race": Race or ethnicity (one of: asian, white, middle eastern, indian, latino, black),
    "gender": Gender (Male or Female),
    "top_color": Color of the top garment (e.g., shirt, blouse),  # Changed from shirt_color
    "bottom_color": Color of the bottom garment (pants, skirt, etc.)
}}

COCOClass Enum:
{{
{', '.join([f'"{member.name}"' for member in COCOClass])}
}}

Important Notes:

1. The output JSON should be as minimal as possible. Do not include fields like 'primary_color' or 'secondary_color' if they are not mentioned in the prompt.

2. Be especially careful with 'activity' and 'complex_appearance' fields. Use them only if the prompt has data that does not fit elsewhere in the JSON. For example:
   - "a guy with red shirt" -> Map 'red shirt' to 'top_color' in PersonProps.
   - "a guy with a black hat" -> Since there isn't any field for 'hat', include "black hat" in 'complex_appearance'.

3. Avoid using 'complex_appearance' and 'activity' fields unless absolutely necessary.

4. Do not include undefined fields or fields not mentioned in the prompt.

5. Use the COCOClass enum for the target class name.

Now, process the following prompt:

'''prompt_text'''

Provide the Investigation JSON object, including only the relevant fields based on the prompt. Do not include any explanations.
"""

# Function to process the prompt
def process_prompt(prompt_text: str, images: List[str], face_images: List[str],
                   system_message: Optional[str] = None, user_message: Optional[str] = None,
                   temperature: float = 0.0) -> Optional[Dict[str, Any]]:
    client = openai.OpenAI(api_key=api_key)

    # Default user message
    if not user_message:
        user_message = ""

    # Prepare messages for the API
    messages = []
    if system_message.strip():
        messages.append({"role": "system", "content": system_message.replace("prompt_text", prompt_text)})
    if user_message.strip():
        messages.append({"role": "user", "content": user_message.replace("prompt_text", prompt_text)})

    response = client.chat.completions.create(
        model="gpt-4o",
        messages=messages,
        response_format={ "type": "json_object" },
        temperature=temperature,
        max_tokens=1000,
    )

    # Extract the content
    content  = response.choices[0].message.content

    # Parse the JSON output
    try:
        investigation_data = json.loads(content)
    except json.JSONDecodeError as e:
        print("Error parsing JSON:", e)
        print("OpenAI response:", content)
        return None

    # Construct the Investigation object
    investigation = parse_investigation(investigation_data, images, face_images)

    # Convert the Investigation object to dictionary
    if investigation:
        investigation_dict = asdict(investigation)
        # Convert enums to their names
        investigation_dict['target'] = investigation.target.name
        if investigation.props:
            if isinstance(investigation.props, VehicleProps) and investigation.props.type:
                investigation_dict['props']['type'] = investigation.props.type.name
            elif isinstance(investigation.props, PersonProps):
                pass  # No enums in PersonProps
        return investigation_dict
    else:
        return None

# Function to parse the Investigation data
def parse_investigation(data: Dict[str, Any], images: List[str], face_images: List[str]) -> Optional[Investigation]:
    # Parse target
    target_name = data.get('target')
    try:
        target_enum = COCOClass[target_name]
    except KeyError:
        print(f"Invalid COCO class name: {target_name}")
        return None

    # Parse activity
    activity_data = data.get('activity')
    if activity_data:
        activity = Activity(
            prompt=activity_data.get('prompt'),
            type=activity_data.get('type')
        )
    else:
        activity = None

    # Parse props
    props_data = data.get('props')
    props = None
    if props_data:
        if 'face_images' in props_data:
            # PersonProps
            props = PersonProps(
                face_images=face_images,
                age=props_data.get('age'),
                race=props_data.get('race'),
                gender=props_data.get('gender'),
                top_color=props_data.get('top_color'),  # Changed from shirt_color
                bottom_color=props_data.get('bottom_color')
            )
        elif 'brand' in props_data:
            # VehicleProps
            vehicle_type_name = props_data.get('type')
            if vehicle_type_name:
                try:
                    vehicle_type_enum = COCOClass[vehicle_type_name]
                except KeyError:
                    print(f"Invalid vehicle type: {vehicle_type_name}")
                    vehicle_type_enum = None
            else:
                vehicle_type_enum = None

            props = VehicleProps(
                brand=props_data.get('brand'),
                type=vehicle_type_enum,
                plate=props_data.get('plate')
            )

    # Construct the Investigation object
    investigation = Investigation(
        target=target_enum,
        images=images,
        activity=activity,
        complex_appearance=data.get('complex_appearance'),
        props=props,
        primary_color=data.get('primary_color'),
        secondary_color=data.get('secondary_color')
    )

    return investigation

# Gradio app
def gradio_app(prompts_text, system_message, user_message, temperature):
    # Split prompts by commas and strip whitespace
    prompts = [p.strip() for p in prompts_text.split('\n') if p.strip()]
    images = ["http://example.com/image1.jpg", "http://example.com/image2.jpg"]
    face_images = ["http://example.com/face1.jpg"]

    results = []
    for p in prompts:
        investigation_dict = process_prompt(
            prompt_text=p,
            images=images,
            face_images=face_images,
            system_message=system_message if system_message else None,
            user_message=user_message if user_message else None,
            temperature=temperature if temperature else 0.0
        )
        results.append(f'{p}\n')
        if investigation_dict:
            results.append(json.dumps(investigation_dict, indent=4))
        else:
            results.append("Failed to process prompt.")

    return "\n\n".join(results)

if __name__ == "__main__":
    # Default values
    default_prompts = "\n".join([
        "A red sports car with a license plate reading 'FAST123'.",
        "An elderly woman wearing a green dress and a pearl necklace.",
        "A cyclist in a yellow jersey riding a blue bicycle.",
        "A group of people playing frisbee in the park.",
        "A man with a large tattoo of a dragon on his right arm.",
        "A black and white cat sitting on a red couch.",
        "A delivery truck with the 'FedEx' logo on the side.",
        "A child holding a red balloon shaped like a dog.",
        "A person wearing a hoodie with the text 'OpenAI' on it.",
        "A woman in a blue swimsuit swimming in the ocean."
    ])

    default_system_message = DEFAULT_SYSTEM_MESSAGE.replace("{{prompt_text}}", "{prompt_text}")  # Prepare for formatting
    default_user_message = ""    # Optional user message
    default_temperature = 0.0    # Default temperature

    # Create Gradio interface
    iface = gr.Interface(
        fn=gradio_app,
        inputs=[
            gr.Textbox(lines=5, label="List of Prompts (comma-separated)", value=default_prompts),
            gr.Textbox(lines=20, label="System Message (optional)", value=default_system_message),
            gr.Textbox(lines=5, label="User Message (optional)", value=default_user_message),
            gr.Slider(minimum=0, maximum=1, step=0.1, label="Temperature", value=default_temperature)
        ],
        outputs="text",
        title="OpenAI Prompt Engineering Tester",
        description="Test different prompts and messages with the OpenAI API."
    )

    # Launch the app
    iface.launch()