webfind 0.5.0

Self-hosted native search engine for AI agents — pure CLI, zero cost
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

use chrono::{DateTime, Utc};
use serde::{Deserialize, Serialize};
use serde_json::json;

use super::request::SearchDepth;

/// Top-level search response
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct SearchResponse {
    pub request_id: String,
    pub query: String,
    pub depth: SearchDepth,
    pub total_results: u64,
    pub returned: u32,
    pub latency_ms: u64,
    pub results: Vec<SearchResult>,
    pub suggestions: Vec<String>,
    pub related: Vec<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub graph: Option<GraphSummary>,
    pub metadata: SearchMetadata,
}

/// Individual search result
/// Classification tier based on threshold outcomes (Funnel Stage 1)
#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize, Default)]
#[serde(rename_all = "snake_case")]
pub enum QualityTier {
    #[default]
    HighQualityRelevant,
    Marginal,
    RejectedLowQuality,
}

/// Extracted practical code, command, or demonstration example from content
#[derive(Debug, Clone, Serialize, Deserialize, PartialEq)]
pub struct ExtractedExample {
    #[serde(skip_serializing_if = "Option::is_none")]
    pub language: Option<String>,
    pub snippet: String,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub context: Option<String>,
    pub relevance: f32,
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct SearchResult {
    pub rank: u32,
    pub url: String,
    pub title: String,
    pub snippet: String,
    pub domain: String,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub published_at: Option<DateTime<Utc>>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub modified_at: Option<DateTime<Utc>>,
    pub crawled_at: DateTime<Utc>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub author: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub site_name: Option<String>,
    pub score: f64,
    pub scores: ScoreBreakdown,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub tier: Option<QualityTier>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub examples: Option<Vec<ExtractedExample>>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub content: Option<ContentBlock>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub keywords: Option<Vec<Keyword>>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub metrics: Option<ContentMetrics>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub favicon: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub thumbnail: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub llms_txt: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub ai_catalog: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub openapi_spec: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub mcp_server: Option<String>,
    pub language: String,
    pub content_type: String,
}

/// Score breakdown by ranking signal
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct ScoreBreakdown {
    #[serde(skip_serializing_if = "Option::is_none")]
    pub bm25: Option<f64>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub vector: Option<f64>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub graph: Option<f64>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub freshness: Option<f64>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub quality: Option<f64>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub ax_score: Option<f64>,
    pub final_score: f64,
}

/// Content block with normalized text
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct ContentBlock {
    pub text: String,
    pub excerpt: String,
    pub word_count: u32,
    pub reading_time_seconds: u32,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub html: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub markdown: Option<String>,
}

/// Extracted keyword
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct Keyword {
    pub text: String,
    pub tfidf_score: f64,
    pub rank: u32,
}

/// Content quality metrics
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct ContentMetrics {
    pub reading_ease: f64,
    pub grade_level: f64,
    pub fog_index: f64,
    pub sentence_count: u32,
    pub avg_words_per_sentence: f64,
    pub language: String,
    pub language_confidence: f64,
    pub has_structured_data: bool,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub schema_type: Option<String>,
}

/// Graph summary for search results
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct GraphSummary {
    pub inbound_links: Vec<GraphEdge>,
    pub outbound_links: Vec<GraphEdge>,
    pub related_domains: Vec<DomainNode>,
    pub domain_authority: Vec<DomainAuthority>,
}

/// Graph edge (link relationship)
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct GraphEdge {
    pub source_url: String,
    pub target_url: String,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub anchor_text: Option<String>,
    pub crawled_at: DateTime<Utc>,
}

/// Domain node in graph
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct DomainNode {
    pub domain: String,
    pub page_count: u32,
    pub avg_authority: f64,
}

/// Domain authority score
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct DomainAuthority {
    pub domain: String,
    pub authority_score: f64,
    pub inbound_links: u32,
    pub outbound_links: u32,
}

/// Search metadata
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct SearchMetadata {
    pub index_version: String,
    pub index_size: u64,
    pub engine_version: String,
    pub searched_at: DateTime<Utc>,
    pub signals_used: Vec<String>,
    pub index_freshness: IndexFreshness,
}

/// Index freshness info
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct IndexFreshness {
    #[serde(skip_serializing_if = "Option::is_none")]
    pub oldest_page: Option<DateTime<Utc>>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub newest_page: Option<DateTime<Utc>>,
    pub avg_age_days: f64,
}

impl SearchResponse {
    /// Render a concise, LLM-friendly view of the response.
    ///
    /// Strips internal score noise and keeps only the information a model needs
    /// to answer a question: title, URL, dates, author, site, excerpt, and —
    /// when requested — the full normalized content block.
    pub fn to_llm_value(&self) -> serde_json::Value {
        let results: Vec<serde_json::Value> = self
            .results
            .iter()
            .map(|r| {
                let mut obj = json!({
                    "rank": r.rank,
                    "title": r.title,
                    "url": r.url,
                    "domain": r.domain,
                    "excerpt": r.snippet,
                    "published_at": r.published_at,
                    "modified_at": r.modified_at,
                    "author": r.author,
                    "site_name": r.site_name,
                    "language": r.language,
                });
                if let Some(ref content) = r.content {
                    obj["word_count"] = json!(content.word_count);
                    obj["reading_time_seconds"] = json!(content.reading_time_seconds);
                    obj["content"] = json!({
                        "text": content.text,
                        "excerpt": content.excerpt,
                        "markdown": content.markdown,
                        "word_count": content.word_count,
                        "reading_time_seconds": content.reading_time_seconds,
                    });
                }
                obj
            })
            .collect();

        let mut value = json!({
            "query": self.query,
            "total_results": self.total_results,
            "returned": self.returned,
            "results": results,
        });

        if self.graph.is_some() {
            value["graph"] = json!(self.graph);
        }

        value["metadata"] = json!({
            "index_size": self.metadata.index_size,
            "engine_version": self.metadata.engine_version,
            "searched_at": self.metadata.searched_at,
            "signals_used": self.metadata.signals_used,
        });

        value
    }
}