cognis 0.3.0

Cognis umbrella crate: agent builder, multi-agent orchestration, memory, middleware (rate limit, retry, PII, prompt caching), built-in tools, and re-exports of cognis-core, cognis-graph, cognis-llm, and cognis-rag.
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

//! Shared test helpers for middleware tests.

use std::sync::atomic::{AtomicUsize, Ordering};
use std::sync::{Arc, Mutex};

use async_trait::async_trait;

use cognis_core::{Message, Result, RunnableStream};
use cognis_llm::chat::{ChatOptions, ChatResponse, HealthStatus, StreamChunk, Usage};
use cognis_llm::provider::{LLMProvider, Provider};

use super::{MiddlewareCtx, Next};

/// Build a canned `ChatResponse` with `text` as the assistant content.
pub(crate) fn ok_resp(text: impl Into<String>) -> ChatResponse {
    ChatResponse {
        message: Message::ai(text),
        usage: None,
        finish_reason: "stop".into(),
        model: "test".into(),
    }
}

/// `Next` that always returns the same canned response.
pub(crate) struct FixedNext(pub ChatResponse);

#[async_trait]
impl Next for FixedNext {
    async fn invoke(&self, _ctx: MiddlewareCtx) -> Result<ChatResponse> {
        Ok(self.0.clone())
    }
}

/// `Next` that records every received ctx then returns a canned response.
pub(crate) struct RecordingNext {
    pub seen: Mutex<Vec<MiddlewareCtx>>,
    pub response: ChatResponse,
}

impl RecordingNext {
    pub(crate) fn new(response: ChatResponse) -> Self {
        Self {
            seen: Mutex::new(Vec::new()),
            response,
        }
    }
}

#[async_trait]
impl Next for RecordingNext {
    async fn invoke(&self, ctx: MiddlewareCtx) -> Result<ChatResponse> {
        self.seen.lock().unwrap().push(ctx);
        Ok(self.response.clone())
    }
}

/// Provider whose response per call is determined by the user-supplied closure.
/// `closure(call_index)` returns either the assistant's text or an error.
pub(crate) struct FlakyProvider {
    counter: AtomicUsize,
    inner: Mutex<Box<dyn Fn(usize) -> Result<String> + Send + Sync>>,
}

pub(crate) fn make_flaky_provider<F>(f: F) -> Arc<FlakyProvider>
where
    F: Fn(usize) -> Result<String> + Send + Sync + 'static,
{
    Arc::new(FlakyProvider {
        counter: AtomicUsize::new(0),
        inner: Mutex::new(Box::new(f)),
    })
}

#[async_trait]
impl LLMProvider for FlakyProvider {
    fn name(&self) -> &str {
        "flaky"
    }
    fn provider_type(&self) -> Provider {
        Provider::Ollama
    }
    async fn chat_completion(
        &self,
        _messages: Vec<Message>,
        _opts: ChatOptions,
    ) -> Result<ChatResponse> {
        let i = self.counter.fetch_add(1, Ordering::SeqCst);
        let f = self.inner.lock().unwrap();
        let text = f(i)?;
        Ok(ChatResponse {
            message: Message::ai(text),
            usage: Some(Usage::default()),
            finish_reason: "stop".into(),
            model: "flaky".into(),
        })
    }
    async fn chat_completion_stream(
        &self,
        _: Vec<Message>,
        _: ChatOptions,
    ) -> Result<RunnableStream<StreamChunk>> {
        unimplemented!()
    }
    async fn health_check(&self) -> Result<HealthStatus> {
        Ok(HealthStatus::Healthy { latency_ms: 0 })
    }
}

/// Provider that captures every received `(messages, opts)` and returns a
/// canned response.
pub(crate) struct RecordingProvider {
    pub received: Mutex<Vec<(Vec<Message>, ChatOptions)>>,
    pub response: Mutex<String>,
}

pub(crate) fn make_recording_provider(
    initial_response: impl Into<String>,
) -> Arc<RecordingProvider> {
    Arc::new(RecordingProvider {
        received: Mutex::new(Vec::new()),
        response: Mutex::new(initial_response.into()),
    })
}

#[async_trait]
impl LLMProvider for RecordingProvider {
    fn name(&self) -> &str {
        "recording"
    }
    fn provider_type(&self) -> Provider {
        Provider::Ollama
    }
    async fn chat_completion(
        &self,
        messages: Vec<Message>,
        opts: ChatOptions,
    ) -> Result<ChatResponse> {
        self.received.lock().unwrap().push((messages, opts));
        Ok(ChatResponse {
            message: Message::ai(self.response.lock().unwrap().clone()),
            usage: Some(Usage::default()),
            finish_reason: "stop".into(),
            model: "recording".into(),
        })
    }
    async fn chat_completion_stream(
        &self,
        _: Vec<Message>,
        _: ChatOptions,
    ) -> Result<RunnableStream<StreamChunk>> {
        unimplemented!()
    }
    async fn health_check(&self) -> Result<HealthStatus> {
        Ok(HealthStatus::Healthy { latency_ms: 0 })
    }
}