octolib 0.4.2

Self-sufficient AI provider library with multi-provider support, embedding models, model validation, and cost tracking
Documentation
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
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
// Copyright 2025 Muvon Un Limited
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

//! Provider strategy pattern for handling provider-specific logic

use crate::errors::{ProviderError, ProviderResult};
use crate::llm::config::CacheConfig;
use crate::llm::tool_calls::ProviderToolCalls;
use crate::llm::types::{Message, ProviderExchange, ToolCall};
use std::collections::HashMap;

/// Strategy trait for provider-specific operations
pub trait ProviderStrategy: Send + Sync {
    /// Get the provider name
    fn provider_name(&self) -> &'static str;

    /// Extract tool calls from provider response
    fn extract_tool_calls(
        &self,
        exchange: &ProviderExchange,
    ) -> ProviderResult<Option<Vec<ToolCall>>>;

    /// Format tool results for sending back to provider
    fn format_tool_results(&self, results: &[ToolResult]) -> ProviderResult<serde_json::Value>;

    /// Apply cache control to message based on provider requirements
    fn apply_cache_control(&self, message: &mut Message, config: &CacheConfig);

    /// Get provider-specific model limits
    fn get_model_limits(&self, model: &str) -> ModelLimits;

    /// Validate model name for this provider
    fn validate_model(&self, model: &str) -> ProviderResult<()>;

    /// Get provider-specific error handling
    fn handle_api_error(&self, status: u16, body: &str) -> ProviderError;
}

/// Tool result for formatting
#[derive(Debug, Clone)]
pub struct ToolResult {
    pub tool_call_id: String,
    pub tool_name: String,
    pub content: String,
    pub is_error: bool,
}

/// Model limits for a provider
#[derive(Debug, Clone)]
pub struct ModelLimits {
    pub max_input_tokens: usize,
    pub max_output_tokens: usize,
    pub supports_vision: bool,
    pub supports_caching: bool,
    pub supports_tools: bool,
}

/// Anthropic provider strategy
pub struct AnthropicStrategy;

impl ProviderStrategy for AnthropicStrategy {
    fn provider_name(&self) -> &'static str {
        "anthropic"
    }

    fn extract_tool_calls(
        &self,
        exchange: &ProviderExchange,
    ) -> ProviderResult<Option<Vec<ToolCall>>> {
        let provider_calls = ProviderToolCalls::extract_from_exchange(exchange)
            .map_err(ProviderError::ToolCallError)?;

        match provider_calls {
            Some(calls) => {
                let tool_calls = calls
                    .to_tool_calls()
                    .map_err(ProviderError::ToolCallError)?;
                Ok(Some(tool_calls))
            }
            None => Ok(None),
        }
    }

    fn format_tool_results(&self, results: &[ToolResult]) -> ProviderResult<serde_json::Value> {
        let content: Vec<serde_json::Value> = results
            .iter()
            .map(|result| {
                serde_json::json!({
                    "type": "tool_result",
                    "tool_use_id": result.tool_call_id,
                    "content": result.content,
                    "is_error": result.is_error
                })
            })
            .collect();

        Ok(serde_json::json!({
            "role": "user",
            "content": content
        }))
    }

    fn apply_cache_control(&self, _message: &mut Message, _config: &CacheConfig) {
        // Anthropic uses cache control in the message content
        // This is handled at the provider level during message conversion
    }

    fn get_model_limits(&self, model: &str) -> ModelLimits {
        if model.contains("claude-opus-4")
            || model.contains("claude-sonnet-4")
            || model.contains("claude-3-5")
        {
            ModelLimits {
                max_input_tokens: 200_000,
                max_output_tokens: 8_192,
                supports_vision: true,
                supports_caching: true,
                supports_tools: true,
            }
        } else if model.contains("claude-3") {
            ModelLimits {
                max_input_tokens: 200_000,
                max_output_tokens: 4_096,
                supports_vision: model.contains("claude-3-opus")
                    || model.contains("claude-3-sonnet"),
                supports_caching: true,
                supports_tools: true,
            }
        } else {
            ModelLimits {
                max_input_tokens: 100_000,
                max_output_tokens: 4_096,
                supports_vision: false,
                supports_caching: false,
                supports_tools: true,
            }
        }
    }

    fn validate_model(&self, model: &str) -> ProviderResult<()> {
        if model.starts_with("claude-") || model.contains("claude") {
            Ok(())
        } else {
            Err(ProviderError::ModelNotSupported {
                provider: "anthropic".to_string(),
                model: model.to_string(),
            })
        }
    }

    fn handle_api_error(&self, status: u16, body: &str) -> ProviderError {
        match status {
            400 => ProviderError::ApiError {
                provider: "anthropic".to_string(),
                status,
                message: format!("Bad Request: {}", body),
            },
            401 => ProviderError::InvalidApiKey {
                provider: "anthropic".to_string(),
            },
            429 => ProviderError::RateLimitExceeded {
                provider: "anthropic".to_string(),
            },
            _ => ProviderError::ApiError {
                provider: "anthropic".to_string(),
                status,
                message: body.to_string(),
            },
        }
    }
}

/// OpenAI provider strategy
pub struct OpenAIStrategy;

impl ProviderStrategy for OpenAIStrategy {
    fn provider_name(&self) -> &'static str {
        "openai"
    }

    fn extract_tool_calls(
        &self,
        exchange: &ProviderExchange,
    ) -> ProviderResult<Option<Vec<ToolCall>>> {
        let provider_calls = ProviderToolCalls::extract_from_exchange(exchange)
            .map_err(ProviderError::ToolCallError)?;

        match provider_calls {
            Some(calls) => {
                let tool_calls = calls
                    .to_tool_calls()
                    .map_err(ProviderError::ToolCallError)?;
                Ok(Some(tool_calls))
            }
            None => Ok(None),
        }
    }

    fn format_tool_results(&self, results: &[ToolResult]) -> ProviderResult<serde_json::Value> {
        let messages: Vec<serde_json::Value> = results
            .iter()
            .map(|result| {
                serde_json::json!({
                    "role": "tool",
                    "tool_call_id": result.tool_call_id,
                    "name": result.tool_name,
                    "content": result.content
                })
            })
            .collect();

        Ok(serde_json::json!(messages))
    }

    fn apply_cache_control(&self, _message: &mut Message, _config: &CacheConfig) {
        // OpenAI doesn't support cache control at the message level
    }

    fn get_model_limits(&self, model: &str) -> ModelLimits {
        if model.contains("gpt-4o") {
            ModelLimits {
                max_input_tokens: 128_000,
                max_output_tokens: 16_384,
                supports_vision: true,
                supports_caching: false,
                supports_tools: true,
            }
        } else if model.contains("gpt-4") {
            ModelLimits {
                max_input_tokens: 128_000,
                max_output_tokens: 4_096,
                supports_vision: model.contains("vision"),
                supports_caching: false,
                supports_tools: true,
            }
        } else if model.contains("gpt-3.5") {
            ModelLimits {
                max_input_tokens: 16_385,
                max_output_tokens: 4_096,
                supports_vision: false,
                supports_caching: false,
                supports_tools: true,
            }
        } else {
            ModelLimits {
                max_input_tokens: 4_096,
                max_output_tokens: 2_048,
                supports_vision: false,
                supports_caching: false,
                supports_tools: false,
            }
        }
    }

    fn validate_model(&self, model: &str) -> ProviderResult<()> {
        if model.starts_with("gpt-") || model.contains("davinci") || model.contains("curie") {
            Ok(())
        } else {
            Err(ProviderError::ModelNotSupported {
                provider: "openai".to_string(),
                model: model.to_string(),
            })
        }
    }

    fn handle_api_error(&self, status: u16, body: &str) -> ProviderError {
        match status {
            400 => ProviderError::ApiError {
                provider: "openai".to_string(),
                status,
                message: format!("Bad Request: {}", body),
            },
            401 => ProviderError::InvalidApiKey {
                provider: "openai".to_string(),
            },
            429 => ProviderError::RateLimitExceeded {
                provider: "openai".to_string(),
            },
            _ => ProviderError::ApiError {
                provider: "openai".to_string(),
                status,
                message: body.to_string(),
            },
        }
    }
}

/// Generic strategy for unknown providers
pub struct GenericStrategy {
    provider_name: String,
}

impl GenericStrategy {
    pub fn new(provider_name: String) -> Self {
        Self { provider_name }
    }
}

impl ProviderStrategy for GenericStrategy {
    fn provider_name(&self) -> &'static str {
        // This is a limitation - we can't return a dynamic string from a static method
        // In practice, this would need to be redesigned or we'd use an enum
        "generic"
    }

    fn extract_tool_calls(
        &self,
        exchange: &ProviderExchange,
    ) -> ProviderResult<Option<Vec<ToolCall>>> {
        let provider_calls = ProviderToolCalls::extract_from_exchange(exchange)
            .map_err(ProviderError::ToolCallError)?;

        match provider_calls {
            Some(calls) => {
                let tool_calls = calls
                    .to_tool_calls()
                    .map_err(ProviderError::ToolCallError)?;
                Ok(Some(tool_calls))
            }
            None => Ok(None),
        }
    }

    fn format_tool_results(&self, results: &[ToolResult]) -> ProviderResult<serde_json::Value> {
        // Generic format - array of tool result objects
        let tool_results: Vec<serde_json::Value> = results
            .iter()
            .map(|result| {
                serde_json::json!({
                    "tool_call_id": result.tool_call_id,
                    "tool_name": result.tool_name,
                    "content": result.content,
                    "is_error": result.is_error
                })
            })
            .collect();

        Ok(serde_json::json!(tool_results))
    }

    fn apply_cache_control(&self, _message: &mut Message, _config: &CacheConfig) {
        // Generic providers don't support cache control
    }

    fn get_model_limits(&self, _model: &str) -> ModelLimits {
        // Conservative defaults for unknown providers
        ModelLimits {
            max_input_tokens: 4_096,
            max_output_tokens: 2_048,
            supports_vision: false,
            supports_caching: false,
            supports_tools: false,
        }
    }

    fn validate_model(&self, _model: &str) -> ProviderResult<()> {
        // Generic strategy accepts any model
        Ok(())
    }

    fn handle_api_error(&self, status: u16, body: &str) -> ProviderError {
        ProviderError::ApiError {
            provider: self.provider_name.clone(),
            status,
            message: body.to_string(),
        }
    }
}

/// Strategy factory for creating appropriate strategies
pub struct StrategyFactory;

impl StrategyFactory {
    /// Get strategy for a provider
    pub fn get_strategy(provider: &str) -> Box<dyn ProviderStrategy> {
        match provider {
            "anthropic" => Box::new(AnthropicStrategy),
            "openai" => Box::new(OpenAIStrategy),
            "openrouter" => Box::new(OpenAIStrategy), // OpenRouter uses OpenAI format
            "deepseek" => Box::new(OpenAIStrategy),   // DeepSeek uses OpenAI format
            _ => Box::new(GenericStrategy::new(provider.to_string())),
        }
    }

    /// Get all available strategies
    pub fn get_all_strategies() -> HashMap<&'static str, Box<dyn ProviderStrategy>> {
        let mut strategies = HashMap::new();
        strategies.insert(
            "anthropic",
            Box::new(AnthropicStrategy) as Box<dyn ProviderStrategy>,
        );
        strategies.insert(
            "openai",
            Box::new(OpenAIStrategy) as Box<dyn ProviderStrategy>,
        );
        strategies
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_strategy_factory() {
        let anthropic_strategy = StrategyFactory::get_strategy("anthropic");
        assert_eq!(anthropic_strategy.provider_name(), "anthropic");

        let openai_strategy = StrategyFactory::get_strategy("openai");
        assert_eq!(openai_strategy.provider_name(), "openai");

        let generic_strategy = StrategyFactory::get_strategy("unknown");
        assert_eq!(generic_strategy.provider_name(), "generic");
    }

    #[test]
    fn test_anthropic_model_validation() {
        let strategy = AnthropicStrategy;

        assert!(strategy.validate_model("claude-3-sonnet").is_ok());
        assert!(strategy.validate_model("claude-opus-4").is_ok());
        assert!(strategy.validate_model("gpt-4").is_err());
    }

    #[test]
    fn test_openai_model_validation() {
        let strategy = OpenAIStrategy;

        assert!(strategy.validate_model("gpt-4o").is_ok());
        assert!(strategy.validate_model("gpt-3.5-turbo").is_ok());
        assert!(strategy.validate_model("claude-3-sonnet").is_err());
    }

    #[test]
    fn test_model_limits() {
        let anthropic_strategy = AnthropicStrategy;
        let limits = anthropic_strategy.get_model_limits("claude-3-5-sonnet");

        assert_eq!(limits.max_input_tokens, 200_000);
        assert!(limits.supports_vision);
        assert!(limits.supports_caching);
        assert!(limits.supports_tools);
    }

    #[test]
    fn test_tool_result_formatting() {
        let anthropic_strategy = AnthropicStrategy;
        let results = vec![ToolResult {
            tool_call_id: "toolu_123".to_string(),
            tool_name: "test_tool".to_string(),
            content: "result content".to_string(),
            is_error: false,
        }];

        let formatted = anthropic_strategy.format_tool_results(&results).unwrap();
        assert_eq!(formatted["role"], "user");
        assert!(formatted["content"].is_array());
    }
}