File size: 7,700 Bytes
ec8f374
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
"""
Synthetic Data Generator Module

Generates synthetic training data using LLM APIs (OpenAI, Anthropic).
"""

import json
import os
from typing import List, Dict, Any, Optional
import time


class SyntheticDataGenerator:
    """Generate synthetic training data using LLMs."""

    def __init__(
        self,
        api_provider: str = "openai",
        api_key: Optional[str] = None,
        model: Optional[str] = None
    ):
        """
        Initialize synthetic data generator.

        Args:
            api_provider: "openai" or "anthropic"
            api_key: API key (uses environment variable if None)
            model: Model name (uses default if None)
        """
        self.api_provider = api_provider.lower()
        self.api_key = api_key or os.getenv("OPENAI_API_KEY" if self.api_provider == "openai" else "ANTHROPIC_API_KEY")

        if self.api_provider == "openai":
            self.model = model or "gpt-4-turbo-preview"
        else:
            self.model = model or "claude-3-opus-20240229"

        self.client = None
        self._init_client()

    def _init_client(self):
        """Initialize API client."""
        try:
            if self.api_provider == "openai":
                from openai import OpenAI
                self.client = OpenAI(api_key=self.api_key)
            else:
                from anthropic import Anthropic
                self.client = Anthropic(api_key=self.api_key)
        except ImportError:
            print(f"Warning: {self.api_provider} library not installed")
        except Exception as e:
            print(f"Warning: Failed to initialize {self.api_provider} client: {e}")

    def generate_examples(
        self,
        num_examples: int,
        topics: List[str],
        difficulty: str = "mixed",
        domain: str = "financial_advisor"
    ) -> List[Dict[str, Any]]:
        """
        Generate synthetic training examples.

        Args:
            num_examples: Number of examples to generate
            topics: List of topics to cover
            difficulty: "beginner", "intermediate", "advanced", or "mixed"
            domain: Domain context

        Returns:
            List of generated examples
        """
        if not self.client:
            raise ValueError(f"API client not initialized. Check {self.api_provider} API key.")

        print(f"Generating {num_examples} examples using {self.api_provider}...")

        examples = []
        examples_per_topic = max(1, num_examples // len(topics))

        for topic in topics:
            for i in range(examples_per_topic):
                if len(examples) >= num_examples:
                    break

                try:
                    example = self._generate_single_example(topic, difficulty, domain)
                    if example:
                        examples.append(example)
                        print(f"Generated {len(examples)}/{num_examples}", end="\r")
                    time.sleep(0.5)  # Rate limiting
                except Exception as e:
                    print(f"\nError generating example: {e}")
                    continue

        print(f"\n✅ Generated {len(examples)} examples")
        return examples

    def _generate_single_example(
        self,
        topic: str,
        difficulty: str,
        domain: str
    ) -> Optional[Dict[str, Any]]:
        """Generate a single training example."""

        prompt = f"""Generate a realistic {domain} training example about {topic}.
Difficulty level: {difficulty}

Output format (JSON):
{{
  "instruction": "The user's question or request",
  "input": "Additional context (optional, can be empty string)",
  "output": "The detailed, helpful response"
}}

Make it realistic and detailed. The response should be informative and professional."""

        try:
            if self.api_provider == "openai":
                response = self.client.chat.completions.create(
                    model=self.model,
                    messages=[
                        {"role": "system", "content": "You are a data generation assistant. Output only valid JSON."},
                        {"role": "user", "content": prompt}
                    ],
                    temperature=0.8
                )
                content = response.choices[0].message.content
            else:  # anthropic
                response = self.client.messages.create(
                    model=self.model,
                    max_tokens=1024,
                    messages=[
                        {"role": "user", "content": prompt}
                    ],
                    temperature=0.8
                )
                content = response.content[0].text

            # Parse JSON
            # Remove markdown code blocks if present
            content = content.strip()
            if content.startswith("```"):
                content = content.split("```")[1]
                if content.startswith("json"):
                    content = content[4:]
            content = content.strip()

            example = json.loads(content)

            # Validate structure
            if "instruction" in example and "output" in example:
                if "input" not in example:
                    example["input"] = ""
                return example

        except Exception as e:
            print(f"\nError parsing example: {e}")

        return None

    def generate_from_scenarios(
        self,
        scenarios: List[Dict[str, Any]],
        num_examples_per_scenario: int = 1
    ) -> List[Dict[str, Any]]:
        """
        Generate examples from predefined scenarios.

        Args:
            scenarios: List of scenario dicts with context
            num_examples_per_scenario: Examples to generate per scenario

        Returns:
            Generated examples
        """
        examples = []

        for scenario in scenarios:
            for _ in range(num_examples_per_scenario):
                prompt = f"""Based on this scenario, generate a training example:

Scenario: {json.dumps(scenario, indent=2)}

Output format (JSON):
{{
  "instruction": "The user's question or request",
  "input": "Additional context based on scenario",
  "output": "The detailed, helpful response"
}}"""

                try:
                    example = self._generate_from_prompt(prompt)
                    if example:
                        examples.append(example)
                except Exception as e:
                    print(f"Error: {e}")
                    continue

        return examples

    def _generate_from_prompt(self, prompt: str) -> Optional[Dict[str, Any]]:
        """Generate example from custom prompt."""
        try:
            if self.api_provider == "openai":
                response = self.client.chat.completions.create(
                    model=self.model,
                    messages=[{"role": "user", "content": prompt}],
                    temperature=0.8
                )
                content = response.choices[0].message.content
            else:
                response = self.client.messages.create(
                    model=self.model,
                    max_tokens=1024,
                    messages=[{"role": "user", "content": prompt}]
                )
                content = response.content[0].text

            # Parse JSON
            content = content.strip()
            if "```" in content:
                content = content.split("```")[1]
                if content.startswith("json"):
                    content = content[4:]
            content = content.strip()

            return json.loads(content)

        except Exception as e:
            print(f"Error: {e}")
            return None