feat: first take at M4

This commit is contained in:
2026-04-10 15:32:05 +02:00
parent ee961f1b02
commit 3ca80b00d0
27 changed files with 3923 additions and 21 deletions

View File

@@ -0,0 +1,756 @@
use std::time::Duration;
use keyring::Entry;
use reqwest::blocking::Client;
use rusqlite::Connection;
use serde::{Deserialize, Serialize};
use serde_json::{json, Value};
use crate::db::queries::setting;
use crate::engine::{EngineError, EngineResult};
use crate::util::now_unix_ms;
const KEYRING_SERVICE: &str = "RuDS";
const KEYRING_SETTING_PREFIX: &str = "ai.endpoint";
#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
#[serde(rename_all = "lowercase")]
pub enum AiEndpointKind {
Online,
Airplane,
}
impl AiEndpointKind {
pub fn as_str(self) -> &'static str {
match self {
Self::Online => "online",
Self::Airplane => "airplane",
}
}
fn settings_prefix(self) -> String {
format!("ai.endpoint.{}", self.as_str())
}
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct AiEndpointConfig {
pub kind: AiEndpointKind,
pub url: String,
pub model: String,
pub api_key: Option<String>,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct StoredAiEndpointConfig {
pub kind: AiEndpointKind,
pub url: String,
pub model: String,
pub api_key_configured: bool,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
pub struct AiSettings {
pub offline_mode: bool,
pub default_model: Option<String>,
pub title_model: Option<String>,
pub image_model: Option<String>,
pub system_prompt: String,
pub online_endpoint: StoredAiEndpointConfig,
pub airplane_endpoint: StoredAiEndpointConfig,
}
impl Default for StoredAiEndpointConfig {
fn default() -> Self {
Self {
kind: AiEndpointKind::Online,
url: String::new(),
model: String::new(),
api_key_configured: false,
}
}
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct AiModelInfo {
pub id: String,
pub name: String,
pub context_window: Option<u64>,
pub max_output_tokens: Option<u64>,
pub supports_vision: bool,
}
#[derive(Debug, Clone, PartialEq, Eq)]
pub enum OneShotOperation {
AnalyzeTaxonomy,
AnalyzePost,
DetectLanguage,
TranslatePost { target_language: String },
AnalyzeImage,
TranslateMedia { target_language: String },
}
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct OneShotRequest {
pub operation: OneShotOperation,
pub content: Value,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct TaxonomySuggestion {
pub tags: Vec<String>,
pub categories: Vec<String>,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct PostAnalysisResult {
pub title: String,
pub excerpt: String,
pub slug: String,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct LanguageDetectionResult {
pub language_code: String,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct TranslationResult {
pub title: String,
pub excerpt: String,
pub content: String,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct ImageAnalysisResult {
pub title: String,
pub alt: String,
pub caption: String,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct MediaTranslationResult {
pub title: String,
pub alt: String,
pub caption: String,
}
#[derive(Debug, Clone, PartialEq, Eq)]
pub enum OneShotResponse {
Taxonomy(TaxonomySuggestion),
PostAnalysis(PostAnalysisResult),
LanguageDetection(LanguageDetectionResult),
Translation(TranslationResult),
ImageAnalysis(ImageAnalysisResult),
MediaTranslation(MediaTranslationResult),
}
pub fn load_ai_settings(conn: &Connection, offline_mode: bool) -> EngineResult<AiSettings> {
let online_endpoint = load_endpoint(conn, AiEndpointKind::Online)?;
let airplane_endpoint = load_endpoint(conn, AiEndpointKind::Airplane)?;
let default_model = get_optional_setting(conn, "ai.default_model")?;
let title_model = get_optional_setting(conn, "ai.title_model")?;
let image_model = get_optional_setting(conn, "ai.image_model")?;
let system_prompt = get_optional_setting(conn, "ai.system_prompt")?.unwrap_or_default();
Ok(AiSettings {
offline_mode,
default_model,
title_model,
image_model,
system_prompt,
online_endpoint,
airplane_endpoint,
})
}
pub fn save_endpoint(conn: &Connection, endpoint: &AiEndpointConfig) -> EngineResult<()> {
validate_endpoint_config(endpoint)?;
let checked_at = now_unix_ms();
set_setting(conn, &endpoint_setting_key(endpoint.kind, "url"), endpoint.url.trim(), checked_at)?;
set_setting(conn, &endpoint_setting_key(endpoint.kind, "model"), endpoint.model.trim(), checked_at)?;
if endpoint.kind == AiEndpointKind::Online {
let entry = endpoint_keyring_entry(endpoint.kind)?;
if let Some(api_key) = &endpoint.api_key {
if api_key.trim().is_empty() {
entry.delete_credential().ok();
set_setting(conn, &endpoint_setting_key(endpoint.kind, "api_key_configured"), "false", checked_at)?;
} else {
entry.set_password(api_key.trim()).map_err(keyring_error)?;
set_setting(conn, &endpoint_setting_key(endpoint.kind, "api_key_configured"), "true", checked_at)?;
}
}
}
Ok(())
}
pub fn save_model_preferences(
conn: &Connection,
default_model: Option<&str>,
title_model: Option<&str>,
image_model: Option<&str>,
system_prompt: &str,
) -> EngineResult<()> {
let updated_at = now_unix_ms();
set_optional_setting(conn, "ai.default_model", default_model, updated_at)?;
set_optional_setting(conn, "ai.title_model", title_model, updated_at)?;
set_optional_setting(conn, "ai.image_model", image_model, updated_at)?;
set_setting(conn, "ai.system_prompt", system_prompt, updated_at)?;
Ok(())
}
pub fn active_endpoint(conn: &Connection, offline_mode: bool) -> EngineResult<AiEndpointConfig> {
let kind = if offline_mode { AiEndpointKind::Airplane } else { AiEndpointKind::Online };
let stored = load_endpoint(conn, kind)?;
if stored.url.trim().is_empty() {
return Err(EngineError::Validation(format!(
"AI unavailable - configure {} endpoint in Settings",
kind.as_str()
)));
}
let api_key = if kind == AiEndpointKind::Online {
let entry = endpoint_keyring_entry(kind)?;
let password = entry.get_password().map_err(keyring_error)?;
if password.trim().is_empty() {
return Err(EngineError::Validation(
"AI unavailable - configure online endpoint in Settings".to_string(),
));
}
Some(password)
} else {
None
};
Ok(AiEndpointConfig {
kind,
url: stored.url,
model: stored.model,
api_key,
})
}
pub fn load_endpoint_api_key(kind: AiEndpointKind) -> EngineResult<Option<String>> {
let entry = endpoint_keyring_entry(kind)?;
match entry.get_password() {
Ok(password) if password.trim().is_empty() => Ok(None),
Ok(password) => Ok(Some(password)),
Err(keyring::Error::NoEntry) => Ok(None),
Err(error) => Err(keyring_error(error)),
}
}
pub fn refresh_model_catalog(endpoint: &AiEndpointConfig) -> EngineResult<Vec<AiModelInfo>> {
validate_endpoint_config(endpoint)?;
let client = build_http_client()?;
let request = client.get(models_url(&endpoint.url));
let response = with_auth(request, endpoint)
.send()?
.error_for_status()?;
let body: Value = response.json()?;
let models = body
.get("data")
.and_then(Value::as_array)
.ok_or_else(|| EngineError::Parse("model catalog response missing data array".to_string()))?;
let mut result = Vec::new();
for model in models {
let id = model
.get("id")
.and_then(Value::as_str)
.ok_or_else(|| EngineError::Parse("model entry missing id".to_string()))?
.to_string();
let name = model
.get("name")
.and_then(Value::as_str)
.unwrap_or(&id)
.to_string();
let context_window = model
.get("context_window")
.or_else(|| model.get("contextWindow"))
.and_then(Value::as_u64);
let max_output_tokens = model
.get("max_output_tokens")
.or_else(|| model.get("maxOutputTokens"))
.and_then(Value::as_u64);
let supports_vision = model
.get("modalities")
.and_then(Value::as_array)
.map(|modalities| modalities.iter().any(|value| value.as_str() == Some("vision")))
.unwrap_or(false);
result.push(AiModelInfo {
id,
name,
context_window,
max_output_tokens,
supports_vision,
});
}
Ok(result)
}
pub fn test_endpoint(endpoint: &AiEndpointConfig) -> EngineResult<()> {
let _ = refresh_model_catalog(endpoint)?;
Ok(())
}
pub fn run_one_shot(
conn: &Connection,
offline_mode: bool,
request: &OneShotRequest,
) -> EngineResult<OneShotResponse> {
let settings = load_ai_settings(conn, offline_mode)?;
let endpoint = active_endpoint(conn, offline_mode)?;
let model = select_model(&settings, &endpoint, &request.operation)?;
let prompt = build_one_shot_prompt(request)?;
let schema = response_schema(&request.operation);
let payload = json!({
"model": model,
"messages": [
{
"role": "system",
"content": build_system_prompt(&settings.system_prompt, &request.operation),
},
{
"role": "user",
"content": prompt,
}
],
"response_format": {
"type": "json_schema",
"json_schema": {
"name": schema.0,
"schema": schema.1,
"strict": true
}
}
});
let client = build_http_client()?;
let response = with_auth(client.post(chat_completions_url(&endpoint.url)).json(&payload), &endpoint)
.send()?
.error_for_status()?;
let body: Value = response.json()?;
let content = body
.get("choices")
.and_then(Value::as_array)
.and_then(|choices| choices.first())
.and_then(|choice| choice.get("message"))
.and_then(|message| message.get("content"))
.and_then(Value::as_str)
.ok_or_else(|| EngineError::Parse("chat completions response missing message content".to_string()))?;
parse_one_shot_response(request, content)
}
fn build_http_client() -> EngineResult<Client> {
Ok(Client::builder().timeout(Duration::from_secs(5)).build()?)
}
fn load_endpoint(conn: &Connection, kind: AiEndpointKind) -> EngineResult<StoredAiEndpointConfig> {
let url = get_optional_setting(conn, &endpoint_setting_key(kind, "url"))?.unwrap_or_default();
let model = get_optional_setting(conn, &endpoint_setting_key(kind, "model"))?.unwrap_or_default();
let api_key_configured = get_optional_setting(conn, &endpoint_setting_key(kind, "api_key_configured"))?
.map(|value| value == "true")
.unwrap_or(false);
Ok(StoredAiEndpointConfig {
kind,
url,
model,
api_key_configured,
})
}
fn validate_endpoint_config(endpoint: &AiEndpointConfig) -> EngineResult<()> {
if endpoint.url.trim().is_empty() {
return Err(EngineError::Validation("endpoint url is required".to_string()));
}
if endpoint.model.trim().is_empty() {
return Err(EngineError::Validation("endpoint model is required".to_string()));
}
if endpoint.kind == AiEndpointKind::Online
&& endpoint.api_key.as_ref().map(|value| value.trim().is_empty()).unwrap_or(true)
{
return Err(EngineError::Validation("online endpoint api key is required".to_string()));
}
Ok(())
}
fn select_model(settings: &AiSettings, endpoint: &AiEndpointConfig, operation: &OneShotOperation) -> EngineResult<String> {
let selected = match operation {
OneShotOperation::AnalyzeImage => settings.image_model.as_ref(),
OneShotOperation::AnalyzeTaxonomy
| OneShotOperation::AnalyzePost
| OneShotOperation::DetectLanguage
| OneShotOperation::TranslatePost { .. }
| OneShotOperation::TranslateMedia { .. } => settings.title_model.as_ref().or(settings.default_model.as_ref()),
}
.filter(|model| !model.trim().is_empty())
.cloned()
.unwrap_or_else(|| endpoint.model.clone());
if selected.trim().is_empty() {
return Err(EngineError::Validation("AI unavailable - configure model in Settings".to_string()));
}
Ok(selected)
}
fn build_system_prompt(base_prompt: &str, operation: &OneShotOperation) -> String {
let operation_prompt = match operation {
OneShotOperation::AnalyzeTaxonomy => "Return only JSON with tags and categories for the post.",
OneShotOperation::AnalyzePost => "Return only JSON with title, excerpt, and slug suggestions for the post.",
OneShotOperation::DetectLanguage => "Return only JSON with the detected language_code.",
OneShotOperation::TranslatePost { target_language } => {
return format!("{} Translate the post into {} and return only JSON with title, excerpt, and content.", base_prompt.trim(), target_language);
}
OneShotOperation::AnalyzeImage => "Return only JSON with title, alt, and caption suggestions for the image.",
OneShotOperation::TranslateMedia { target_language } => {
return format!("{} Translate the media metadata into {} and return only JSON with title, alt, and caption.", base_prompt.trim(), target_language);
}
};
if base_prompt.trim().is_empty() {
operation_prompt.to_string()
} else {
format!("{} {}", base_prompt.trim(), operation_prompt)
}
}
fn build_one_shot_prompt(request: &OneShotRequest) -> EngineResult<String> {
match &request.operation {
OneShotOperation::AnalyzeTaxonomy => Ok(format!(
"Suggest tags and categories for this post: {}",
serde_json::to_string(&request.content)?
)),
OneShotOperation::AnalyzePost => Ok(format!(
"Analyze this post and suggest title, excerpt, and slug: {}",
serde_json::to_string(&request.content)?
)),
OneShotOperation::DetectLanguage => Ok(format!(
"Detect the language of this text: {}",
serde_json::to_string(&request.content)?
)),
OneShotOperation::TranslatePost { target_language } => Ok(format!(
"Translate this post to {}: {}",
target_language,
serde_json::to_string(&request.content)?
)),
OneShotOperation::AnalyzeImage => Ok(format!(
"Analyze this image metadata and return title, alt, and caption suggestions: {}",
serde_json::to_string(&request.content)?
)),
OneShotOperation::TranslateMedia { target_language } => Ok(format!(
"Translate this media metadata to {}: {}",
target_language,
serde_json::to_string(&request.content)?
)),
}
}
fn response_schema(operation: &OneShotOperation) -> (&'static str, Value) {
match operation {
OneShotOperation::AnalyzeTaxonomy => (
"taxonomy_suggestion",
json!({
"type": "object",
"additionalProperties": false,
"properties": {
"tags": { "type": "array", "items": { "type": "string" } },
"categories": { "type": "array", "items": { "type": "string" } }
},
"required": ["tags", "categories"]
}),
),
OneShotOperation::AnalyzePost => (
"post_analysis",
json!({
"type": "object",
"additionalProperties": false,
"properties": {
"title": { "type": "string" },
"excerpt": { "type": "string" },
"slug": { "type": "string" }
},
"required": ["title", "excerpt", "slug"]
}),
),
OneShotOperation::DetectLanguage => (
"language_detection",
json!({
"type": "object",
"additionalProperties": false,
"properties": {
"language_code": { "type": "string" }
},
"required": ["language_code"]
}),
),
OneShotOperation::TranslatePost { .. } => (
"post_translation",
json!({
"type": "object",
"additionalProperties": false,
"properties": {
"title": { "type": "string" },
"excerpt": { "type": "string" },
"content": { "type": "string" }
},
"required": ["title", "excerpt", "content"]
}),
),
OneShotOperation::AnalyzeImage => (
"image_analysis",
json!({
"type": "object",
"additionalProperties": false,
"properties": {
"title": { "type": "string" },
"alt": { "type": "string" },
"caption": { "type": "string" }
},
"required": ["title", "alt", "caption"]
}),
),
OneShotOperation::TranslateMedia { .. } => (
"media_translation",
json!({
"type": "object",
"additionalProperties": false,
"properties": {
"title": { "type": "string" },
"alt": { "type": "string" },
"caption": { "type": "string" }
},
"required": ["title", "alt", "caption"]
}),
),
}
}
fn parse_one_shot_response(request: &OneShotRequest, content: &str) -> EngineResult<OneShotResponse> {
Ok(match request.operation {
OneShotOperation::AnalyzeTaxonomy => OneShotResponse::Taxonomy(serde_json::from_str(content)?),
OneShotOperation::AnalyzePost => OneShotResponse::PostAnalysis(serde_json::from_str(content)?),
OneShotOperation::DetectLanguage => OneShotResponse::LanguageDetection(serde_json::from_str(content)?),
OneShotOperation::TranslatePost { .. } => OneShotResponse::Translation(serde_json::from_str(content)?),
OneShotOperation::AnalyzeImage => OneShotResponse::ImageAnalysis(serde_json::from_str(content)?),
OneShotOperation::TranslateMedia { .. } => OneShotResponse::MediaTranslation(serde_json::from_str(content)?),
})
}
fn endpoint_setting_key(kind: AiEndpointKind, suffix: &str) -> String {
format!("{}.{}", kind.settings_prefix(), suffix)
}
fn endpoint_keyring_entry(kind: AiEndpointKind) -> EngineResult<Entry> {
Entry::new(KEYRING_SERVICE, &format!("{}.{}", KEYRING_SETTING_PREFIX, kind.as_str())).map_err(keyring_error)
}
fn keyring_error(error: keyring::Error) -> EngineError {
EngineError::Validation(error.to_string())
}
fn set_setting(conn: &Connection, key: &str, value: &str, updated_at: i64) -> EngineResult<()> {
setting::set_setting_value(conn, key, value, updated_at)?;
Ok(())
}
fn set_optional_setting(conn: &Connection, key: &str, value: Option<&str>, updated_at: i64) -> EngineResult<()> {
set_setting(conn, key, value.unwrap_or(""), updated_at)
}
fn get_optional_setting(conn: &Connection, key: &str) -> EngineResult<Option<String>> {
match setting::get_setting_by_key(conn, key) {
Ok(setting) if setting.value.trim().is_empty() => Ok(None),
Ok(setting) => Ok(Some(setting.value)),
Err(rusqlite::Error::QueryReturnedNoRows) => Ok(None),
Err(error) => Err(EngineError::Db(error)),
}
}
fn models_url(base_url: &str) -> String {
join_openai_path(base_url, "models")
}
fn chat_completions_url(base_url: &str) -> String {
join_openai_path(base_url, "chat/completions")
}
fn join_openai_path(base_url: &str, suffix: &str) -> String {
let trimmed = base_url.trim_end_matches('/');
if trimmed.ends_with("/v1") {
format!("{trimmed}/{suffix}")
} else {
format!("{trimmed}/v1/{suffix}")
}
}
fn with_auth(
request: reqwest::blocking::RequestBuilder,
endpoint: &AiEndpointConfig,
) -> reqwest::blocking::RequestBuilder {
if let Some(api_key) = &endpoint.api_key {
request.bearer_auth(api_key)
} else {
request
}
}
#[cfg(test)]
mod tests {
use std::io::{Read, Write};
use std::net::TcpListener;
use std::thread;
use super::*;
use crate::db::Database;
fn setup() -> Database {
let mut db = Database::open_in_memory().unwrap();
db.migrate().unwrap();
db
}
fn clear_keyring(kind: AiEndpointKind) {
let entry = endpoint_keyring_entry(kind).unwrap();
entry.delete_credential().ok();
}
#[test]
fn loads_empty_defaults() {
clear_keyring(AiEndpointKind::Online);
let db = setup();
let settings = load_ai_settings(db.conn(), false).unwrap();
assert!(!settings.offline_mode);
assert!(settings.online_endpoint.url.is_empty());
assert!(settings.airplane_endpoint.url.is_empty());
assert!(settings.default_model.is_none());
}
#[test]
fn saves_online_endpoint_with_keychain_secret() {
clear_keyring(AiEndpointKind::Online);
let db = setup();
save_endpoint(
db.conn(),
&AiEndpointConfig {
kind: AiEndpointKind::Online,
url: "https://example.test/v1".to_string(),
model: "gpt-4.1-mini".to_string(),
api_key: Some("secret-token".to_string()),
},
)
.unwrap();
let active = active_endpoint(db.conn(), false).unwrap();
assert_eq!(active.url, "https://example.test/v1");
assert_eq!(active.model, "gpt-4.1-mini");
assert_eq!(active.api_key.as_deref(), Some("secret-token"));
}
#[test]
fn airplane_endpoint_does_not_require_api_key() {
let db = setup();
save_endpoint(
db.conn(),
&AiEndpointConfig {
kind: AiEndpointKind::Airplane,
url: "http://localhost:11434/v1".to_string(),
model: "llama3.2".to_string(),
api_key: None,
},
)
.unwrap();
let active = active_endpoint(db.conn(), true).unwrap();
assert_eq!(active.kind, AiEndpointKind::Airplane);
assert!(active.api_key.is_none());
}
#[test]
fn refresh_model_catalog_parses_openai_models_shape() {
let server = spawn_test_server(|request| {
assert!(request.starts_with("GET /v1/models HTTP/1.1"));
http_ok(
r#"{"data":[{"id":"gpt-4.1-mini","name":"GPT 4.1 mini","context_window":128000,"max_output_tokens":8192,"modalities":["text"]},{"id":"gpt-4.1","modalities":["text","vision"]}]}"#,
)
});
let models = refresh_model_catalog(&AiEndpointConfig {
kind: AiEndpointKind::Airplane,
url: server,
model: "gpt-4.1-mini".to_string(),
api_key: None,
})
.unwrap();
assert_eq!(models.len(), 2);
assert_eq!(models[0].name, "GPT 4.1 mini");
assert!(models[1].supports_vision);
}
#[test]
fn run_one_shot_uses_active_endpoint_and_parses_response() {
clear_keyring(AiEndpointKind::Online);
let server = spawn_test_server(|request| {
if request.starts_with("GET /v1/models HTTP/1.1") {
return http_ok(r#"{"data":[{"id":"gpt-4.1-mini"}]}"#);
}
assert!(request.starts_with("POST /v1/chat/completions HTTP/1.1"));
assert!(request.contains("authorization: Bearer secret-token") || request.contains("Authorization: Bearer secret-token"));
http_ok(
r#"{"choices":[{"message":{"content":"{\"title\":\"Better title\",\"excerpt\":\"Short summary\",\"slug\":\"better-title\"}"}}]}"#,
)
});
let db = setup();
save_endpoint(
db.conn(),
&AiEndpointConfig {
kind: AiEndpointKind::Online,
url: server,
model: "gpt-4.1-mini".to_string(),
api_key: Some("secret-token".to_string()),
},
)
.unwrap();
save_model_preferences(db.conn(), None, Some("gpt-4.1-mini"), None, "").unwrap();
let response = run_one_shot(
db.conn(),
false,
&OneShotRequest {
operation: OneShotOperation::AnalyzePost,
content: json!({"title":"Draft title","excerpt":"","content":"Body"}),
},
)
.unwrap();
assert_eq!(
response,
OneShotResponse::PostAnalysis(PostAnalysisResult {
title: "Better title".to_string(),
excerpt: "Short summary".to_string(),
slug: "better-title".to_string(),
})
);
}
fn spawn_test_server(handler: impl Fn(String) -> String + Send + 'static) -> String {
let listener = TcpListener::bind("127.0.0.1:0").unwrap();
let addr = listener.local_addr().unwrap();
thread::spawn(move || {
for stream in listener.incoming().take(2) {
let mut stream = stream.unwrap();
let mut buffer = [0_u8; 8192];
let size = stream.read(&mut buffer).unwrap();
let request = String::from_utf8_lossy(&buffer[..size]).to_string();
let response = handler(request);
stream.write_all(response.as_bytes()).unwrap();
}
});
format!("http://{}", addr)
}
fn http_ok(body: &str) -> String {
format!(
"HTTP/1.1 200 OK\r\ncontent-type: application/json\r\ncontent-length: {}\r\nconnection: close\r\n\r\n{}",
body.len(),
body
)
}
}

View File

@@ -46,6 +46,12 @@ impl From<std::io::Error> for EngineError {
}
}
impl From<reqwest::Error> for EngineError {
fn from(e: reqwest::Error) -> Self {
Self::Parse(e.to_string())
}
}
impl From<serde_json::Error> for EngineError {
fn from(e: serde_json::Error) -> Self {
Self::Parse(e.to_string())

View File

@@ -0,0 +1,227 @@
use std::path::Path;
use chrono::{DateTime, TimeZone, Utc};
use rusqlite::Connection;
use crate::engine::{EngineError, EngineResult};
use crate::model::{Post, ProjectMetadata};
use crate::render::{
GeneratedWriteOutcome, build_calendar_json, build_canonical_post_path,
render_markdown_to_html, render_starter_list_page, render_starter_single_post_page,
write_generated_file,
};
#[derive(Debug, Clone)]
pub struct PublishedPostSource {
pub post: Post,
pub body_markdown: String,
}
#[derive(Debug, Default, Clone)]
pub struct GenerationReport {
pub written_paths: Vec<String>,
pub skipped_paths: Vec<String>,
}
pub fn generate_starter_site(
conn: &Connection,
output_dir: &Path,
project_id: &str,
metadata: &ProjectMetadata,
posts: &[PublishedPostSource],
language: &str,
) -> EngineResult<GenerationReport> {
let mut report = GenerationReport::default();
let list_input = posts
.iter()
.map(|source| (source.post.clone(), source.body_markdown.clone()))
.collect::<Vec<_>>();
let index_page = render_starter_list_page(&list_input, metadata, language)
.map_err(|error| EngineError::Parse(error.to_string()))?;
write_out(conn, output_dir, project_id, &index_page.relative_path, &index_page.html, &mut report)?;
for source in posts {
let rendered = render_starter_single_post_page(&source.post, &source.body_markdown, metadata, language)
.map_err(|error| EngineError::Parse(error.to_string()))?;
write_out(conn, output_dir, project_id, &rendered.relative_path, &rendered.html, &mut report)?;
}
write_out(
conn,
output_dir,
project_id,
"calendar.json",
&build_calendar_json(&posts.iter().map(|source| source.post.clone()).collect::<Vec<_>>())?,
&mut report,
)?;
let rss = build_rss_xml(metadata, posts, language);
write_out(conn, output_dir, project_id, "rss.xml", &rss, &mut report)?;
write_out(conn, output_dir, project_id, "feed.xml", &rss, &mut report)?;
write_out(conn, output_dir, project_id, "atom.xml", &build_atom_xml(metadata, posts, language), &mut report)?;
write_out(conn, output_dir, project_id, "sitemap.xml", &build_sitemap_xml(metadata, posts, language), &mut report)?;
Ok(report)
}
fn write_out(
conn: &Connection,
output_dir: &Path,
project_id: &str,
relative_path: &str,
content: &str,
report: &mut GenerationReport,
) -> EngineResult<()> {
match write_generated_file(conn, output_dir, project_id, relative_path, content)
.map_err(|error| EngineError::Parse(error.to_string()))?
{
GeneratedWriteOutcome::Written => report.written_paths.push(relative_path.to_string()),
GeneratedWriteOutcome::SkippedUnchanged => report.skipped_paths.push(relative_path.to_string()),
}
Ok(())
}
fn build_rss_xml(metadata: &ProjectMetadata, posts: &[PublishedPostSource], language: &str) -> String {
let base_url = metadata.public_url.as_deref().unwrap_or("").trim_end_matches('/');
let last_build = posts
.iter()
.filter_map(|post| timestamp(post.post.published_at.unwrap_or(post.post.created_at)))
.max()
.unwrap_or_else(Utc::now);
let mut xml = vec![
"<?xml version=\"1.0\" encoding=\"UTF-8\"?>".to_string(),
"<rss version=\"2.0\" xmlns:content=\"http://purl.org/rss/1.0/modules/content/\" xmlns:dc=\"http://purl.org/dc/elements/1.1/\">".to_string(),
" <channel>".to_string(),
format!(" <title>{}</title>", escape_xml(&metadata.name)),
format!(" <link>{base_url}/</link>"),
format!(" <description>{}</description>", escape_xml(metadata.description.as_deref().unwrap_or(""))),
format!(" <lastBuildDate>{}</lastBuildDate>", last_build.format("%a, %d %b %Y %H:%M:%S GMT")),
" <generator>bDS</generator>".to_string(),
];
for source in posts {
let url = format!("{base_url}{}", build_canonical_post_path(&source.post, language, metadata.main_language.as_deref().unwrap_or("en")));
let published = timestamp(source.post.published_at.unwrap_or(source.post.created_at)).unwrap_or_else(Utc::now);
xml.push(" <item>".to_string());
xml.push(format!(" <title>{}</title>", escape_xml(&source.post.title)));
xml.push(format!(" <link>{url}</link>"));
xml.push(format!(" <guid isPermaLink=\"true\">{url}</guid>"));
xml.push(format!(" <pubDate>{}</pubDate>", published.format("%a, %d %b %Y %H:%M:%S GMT")));
if let Some(author) = &source.post.author {
xml.push(format!(" <author>{}</author>", escape_xml(author)));
}
xml.push(format!(" <dc:language>{}</dc:language>", escape_xml(language)));
let html = render_markdown_to_html(&source.body_markdown);
xml.push(format!(" <description><![CDATA[{html}]]></description>"));
xml.push(format!(" <content:encoded><![CDATA[{html}]]></content:encoded>"));
for category in &source.post.categories {
xml.push(format!(" <category>{}</category>", escape_xml(category)));
}
for tag in &source.post.tags {
xml.push(format!(" <category>{}</category>", escape_xml(tag)));
}
xml.push(" </item>".to_string());
}
xml.push(" </channel>".to_string());
xml.push("</rss>".to_string());
xml.join("\n")
}
fn build_atom_xml(metadata: &ProjectMetadata, posts: &[PublishedPostSource], language: &str) -> String {
let base_url = metadata.public_url.as_deref().unwrap_or("").trim_end_matches('/');
let updated = posts
.iter()
.filter_map(|post| timestamp(post.post.published_at.unwrap_or(post.post.created_at)))
.max()
.unwrap_or_else(Utc::now);
let mut xml = vec![
"<?xml version=\"1.0\" encoding=\"UTF-8\"?>".to_string(),
"<feed xmlns=\"http://www.w3.org/2005/Atom\">".to_string(),
format!(" <title>{}</title>", escape_xml(&metadata.name)),
format!(" <subtitle>{}</subtitle>", escape_xml(metadata.description.as_deref().unwrap_or(""))),
format!(" <id>{base_url}/</id>"),
format!(" <link href=\"{base_url}/\" rel=\"alternate\" />"),
format!(" <link href=\"{base_url}/atom.xml\" rel=\"self\" />"),
format!(" <updated>{}</updated>", updated.to_rfc3339_opts(chrono::SecondsFormat::Millis, true)),
];
for source in posts {
let url = format!("{base_url}{}", build_canonical_post_path(&source.post, language, metadata.main_language.as_deref().unwrap_or("en")));
let published = timestamp(source.post.published_at.unwrap_or(source.post.created_at)).unwrap_or_else(Utc::now);
let html = render_markdown_to_html(&source.body_markdown);
xml.push(format!(" <entry xml:lang=\"{}\">", escape_xml(language)));
xml.push(format!(" <title>{}</title>", escape_xml(&source.post.title)));
xml.push(format!(" <id>{url}</id>"));
xml.push(format!(" <link href=\"{url}\" />"));
xml.push(format!(" <updated>{}</updated>", published.to_rfc3339_opts(chrono::SecondsFormat::Millis, true)));
xml.push(format!(" <published>{}</published>", published.to_rfc3339_opts(chrono::SecondsFormat::Millis, true)));
if let Some(author) = &source.post.author {
xml.push(format!(" <author><name>{}</name></author>", escape_xml(author)));
}
xml.push(format!(" <summary type=\"xhtml\"><div xmlns=\"http://www.w3.org/1999/xhtml\">{html}</div></summary>"));
xml.push(format!(" <content type=\"xhtml\"><div xmlns=\"http://www.w3.org/1999/xhtml\">{html}</div></content>"));
for category in &source.post.categories {
xml.push(format!(" <category term=\"{}\" />", escape_xml(category)));
}
for tag in &source.post.tags {
xml.push(format!(" <category term=\"{}\" />", escape_xml(tag)));
}
xml.push(" </entry>".to_string());
}
xml.push("</feed>".to_string());
xml.join("\n")
}
fn build_sitemap_xml(metadata: &ProjectMetadata, posts: &[PublishedPostSource], language: &str) -> String {
let base_url = metadata.public_url.as_deref().unwrap_or("").trim_end_matches('/');
let index_lastmod = posts
.iter()
.filter_map(|post| timestamp(post.post.published_at.unwrap_or(post.post.created_at)))
.max()
.unwrap_or_else(Utc::now)
.to_rfc3339_opts(chrono::SecondsFormat::Millis, true);
let mut xml = vec![
"<?xml version=\"1.0\" encoding=\"UTF-8\"?>".to_string(),
"<urlset xmlns=\"http://www.sitemaps.org/schemas/sitemap/0.9\" xmlns:xhtml=\"http://www.w3.org/1999/xhtml\">".to_string(),
" <url>".to_string(),
format!(" <loc>{base_url}/</loc>"),
format!(" <lastmod>{index_lastmod}</lastmod>"),
" <changefreq>daily</changefreq>".to_string(),
" <priority>1.0</priority>".to_string(),
" </url>".to_string(),
];
for source in posts {
let url = format!("{base_url}{}", build_canonical_post_path(&source.post, language, metadata.main_language.as_deref().unwrap_or("en")));
let lastmod = timestamp(source.post.published_at.unwrap_or(source.post.created_at))
.unwrap_or_else(Utc::now)
.to_rfc3339_opts(chrono::SecondsFormat::Millis, true);
xml.push(" <url>".to_string());
xml.push(format!(" <loc>{url}</loc>"));
xml.push(format!(" <lastmod>{lastmod}</lastmod>"));
xml.push(" <changefreq>weekly</changefreq>".to_string());
xml.push(" <priority>0.8</priority>".to_string());
xml.push(" </url>".to_string());
}
xml.push("</urlset>".to_string());
xml.join("\n")
}
fn timestamp(timestamp_ms: i64) -> Option<DateTime<Utc>> {
chrono::Utc.timestamp_millis_opt(timestamp_ms).single()
}
fn escape_xml(value: &str) -> String {
value
.replace('&', "&amp;")
.replace('<', "&lt;")
.replace('>', "&gt;")
.replace('"', "&quot;")
.replace('\'', "&apos;")
}

View File

@@ -1,4 +1,5 @@
pub mod error;
pub mod ai;
pub mod context;
pub mod project;
pub mod meta;
@@ -15,6 +16,8 @@ pub mod metadata_diff;
pub mod rebuild;
pub mod search;
pub mod calendar;
pub mod generation;
pub mod preview;
pub mod validate_translations;
pub mod validate_media;
pub mod validate_content;

View File

@@ -0,0 +1,126 @@
use crate::engine::{EngineError, EngineResult};
use crate::engine::generation::PublishedPostSource;
use crate::model::ProjectMetadata;
use crate::render::{build_canonical_post_path, render_starter_list_page, render_starter_single_post_page};
pub fn render_preview_path(
path: &str,
metadata: &ProjectMetadata,
posts: &[PublishedPostSource],
) -> EngineResult<Option<String>> {
let normalized = if path.is_empty() { "/" } else { path };
let main_language = metadata.main_language.as_deref().unwrap_or("en");
if normalized == "/" {
let list_posts = posts
.iter()
.map(|source| (source.post.clone(), source.body_markdown.clone()))
.collect::<Vec<_>>();
return render_starter_list_page(&list_posts, metadata, main_language)
.map(|page| Some(page.html))
.map_err(|error| EngineError::Parse(error.to_string()));
}
let (language, route_path) = split_language_prefix(normalized, metadata);
if let Some(source) = posts.iter().find(|source| {
build_canonical_post_path(&source.post, &language, main_language) == route_path
}) {
return render_starter_single_post_page(&source.post, &source.body_markdown, metadata, &language)
.map(|page| Some(page.html))
.map_err(|error| EngineError::Parse(error.to_string()));
}
Ok(None)
}
fn split_language_prefix(path: &str, metadata: &ProjectMetadata) -> (String, String) {
let trimmed = path.trim_start_matches('/');
let mut segments = trimmed.split('/');
let first = segments.next().unwrap_or_default();
if metadata.blog_languages.iter().any(|language| language == first) {
let remainder = segments.collect::<Vec<_>>().join("/");
return (first.to_string(), format!("/{first}/{}", remainder.trim_start_matches('/')));
}
(
metadata.main_language.as_deref().unwrap_or("en").to_string(),
path.to_string(),
)
}
#[cfg(test)]
mod tests {
use super::*;
use crate::model::{Post, PostStatus};
fn make_metadata() -> ProjectMetadata {
ProjectMetadata {
name: "Blog".into(),
description: None,
public_url: Some("https://example.com".into()),
main_language: Some("en".into()),
default_author: None,
max_posts_per_page: 50,
blogmark_category: None,
pico_theme: None,
semantic_similarity_enabled: false,
blog_languages: vec!["en".into(), "de".into()],
}
}
fn make_post() -> PublishedPostSource {
PublishedPostSource {
post: Post {
id: "post-1".into(),
project_id: "project-1".into(),
title: "Hello".into(),
slug: "hello".into(),
excerpt: None,
content: Some("Body".into()),
status: PostStatus::Published,
author: None,
language: Some("en".into()),
do_not_translate: false,
template_slug: None,
file_path: String::new(),
checksum: None,
tags: vec![],
categories: vec![],
published_title: None,
published_content: None,
published_tags: None,
published_categories: None,
published_excerpt: None,
created_at: 1_710_000_000_000,
updated_at: 1_710_000_000_000,
published_at: Some(1_710_000_000_000),
},
body_markdown: "Hello **world**".into(),
}
}
#[test]
fn root_preview_renders_index_page() {
let html = render_preview_path("/", &make_metadata(), &[make_post()])
.unwrap()
.unwrap();
assert!(html.contains("post-list"));
}
#[test]
fn preview_renders_single_post_for_canonical_path() {
let html = render_preview_path("/2024/03/09/hello", &make_metadata(), &[make_post()])
.unwrap()
.unwrap();
assert!(html.contains("<h1>Hello</h1>"));
assert!(html.contains("<strong>world</strong>"));
}
#[test]
fn preview_renders_language_prefixed_single_post() {
let html = render_preview_path("/de/2024/03/09/hello", &make_metadata(), &[make_post()])
.unwrap()
.unwrap();
assert!(html.contains("lang=\"de\""));
}
}