tyrell 0.1.2

More human than human. Interacting with LLMs in a rusty way.
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

use anyhow::Result;
use jsonxf::pretty_print;
use schemars::JsonSchema;
use serde::{Deserialize, Serialize};
use std::fs;
use tyrell::{ClaudeRequest, ContentType, Model, Role, Tool, ToolBuilder, ToolChoice};

#[derive(Debug, Serialize, Deserialize, JsonSchema)]
pub struct EarningsCallAnalysis {
    /// Company name
    company_name: String,
    /// Stock ticker symbol
    ticker: String,
    /// Date of the earnings call
    call_date: String,
    /// Fiscal quarter and year (e.g., "Q2 2023")
    fiscal_period: String,
    /// Reported earnings per share (EPS)
    reported_eps: f64,
    /// Analyst consensus EPS estimate
    estimated_eps: f64,
    /// Reported revenue
    reported_revenue: f64,
    /// Analyst consensus revenue estimate
    estimated_revenue: f64,
    /// Year-over-year revenue growth rate
    yoy_revenue_growth: f64,
    /// Net income for the quarter
    net_income: f64,
    /// Key performance indicators (KPIs) mentioned in the call
    kpis: Vec<KPI>,
    /// Notable quotes from executives
    key_quotes: Vec<String>,
    /// Forward-looking statements or guidance
    guidance: Option<Guidance>,
    /// Major announcements or updates
    announcements: Vec<String>,
    /// Sentiment analysis of the call
    sentiment: CallSentiment,
    /// Potential risk factors mentioned
    risk_factors: Vec<String>,
    /// Analyst questions and management responses
    qa_summary: Vec<QAItem>,
}

#[derive(Debug, Serialize, Deserialize, JsonSchema)]
pub struct KPI {
    /// Name of the KPI
    name: String,
    /// Value of the KPI
    value: String,
    /// Previous period's value, if mentioned
    previous_value: Option<String>,
}

#[derive(Debug, Serialize, Deserialize, JsonSchema)]
pub struct Guidance {
    /// Expected revenue range for next quarter
    next_quarter_revenue: Option<(f64, f64)>,
    /// Expected EPS range for next quarter
    next_quarter_eps: Option<(f64, f64)>,
    /// Expected revenue range for full year
    full_year_revenue: Option<(f64, f64)>,
    /// Expected EPS range for full year
    full_year_eps: Option<(f64, f64)>,
}

#[derive(Debug, Serialize, Deserialize, JsonSchema)]
pub struct CallSentiment {
    /// Overall sentiment score (-1.0 to 1.0)
    overall_score: f64,
    /// Sentiment towards company performance
    performance_sentiment: String,
    /// Sentiment towards future outlook
    outlook_sentiment: String,
}

#[derive(Debug, Serialize, Deserialize, JsonSchema)]
pub struct QAItem {
    /// Question asked by the analyst
    question: String,
    /// Summary of management's response
    response_summary: String,
}

impl ToolBuilder for EarningsCallAnalysis {
    fn name() -> &'static str {
        "analyze_earnings_call"
    }

    fn description() -> Option<&'static str> {
        Some("Extract information from a quarterly earnings call")
    }
}

#[tokio::main]
async fn main() -> Result<()> {
    let tool = Tool::new::<EarningsCallAnalysis>();

    // https://s2.q4cdn.com/661678649/files/doc_financials/2024/q2/2Q24-Boeing-Earnings-Call-Transcript.pdf
    let earnings_call_transcript =
        fs::read_to_string("examples/boeing_2024q2_earnings_transcript.txt")?;

    let chat = ClaudeRequest::builder()
        .model(Model::Haiku3)
        .add_message(
            Role::Assistant,
            vec![ContentType::Text {
                text: "You are an expert financial analyst.".to_string(),
            }],
        )
        .add_message(
            Role::User,
            vec![ContentType::Text {
                text: format!(
                    "Analyze the this quarterly earnings call:\n\n{}",
                    earnings_call_transcript
                ),
            }],
        )
        .max_tokens(200)
        .tools(vec![tool])
        .tool_choice(ToolChoice::Specific {
            // TODO: should name be checked that it matches
            // the tool?
            name: "analyze_earnings_call".to_string(),
            disable_parallel_tool_use: Some(false),
        })
        .build()
        .unwrap();

    let response = chat.call().await.unwrap();
    let response = pretty_print(&response).unwrap();

    println!("{}", response);

    Ok(())
}