Struct ContentGenerator 

pub struct ContentGenerator { /* private fields */ }

Content generator that combines an LLM provider with business context.

Implementations

impl ContentGenerator

pub fn new(provider: Box<dyn LlmProvider>, business: BusinessProfile) -> Self

Create a new content generator.

pub async fn generate_reply( &self, tweet_text: &str, tweet_author: &str, mention_product: bool, ) -> Result<String, LlmError>

Generate a reply to a tweet.

The reply will be conversational, helpful, and under 280 characters. When mention_product is false, the system prompt explicitly forbids mentioning the product name. When archetype is provided, the prompt includes archetype-specific guidance for varied output (e.g., ask a question, share experience). Retries once with a stricter prompt if the first attempt is too long, then truncates at a sentence boundary as a last resort.

pub async fn generate_reply_with_archetype( &self, tweet_text: &str, tweet_author: &str, mention_product: bool, archetype: Option<ReplyArchetype>, ) -> Result<String, LlmError>

Generate a reply using a specific archetype for varied output.

pub async fn generate_tweet(&self, topic: &str) -> Result<String, LlmError>

Generate a standalone educational tweet.

The tweet will be informative, engaging, and under 280 characters.

pub async fn generate_tweet_with_format( &self, topic: &str, format: Option<TweetFormat>, ) -> Result<String, LlmError>

Generate a tweet using a specific format for varied structure.

pub async fn generate_thread( &self, topic: &str, ) -> Result<Vec<String>, LlmError>

Generate an educational thread of 5-8 tweets.

Each tweet in the thread will be under 280 characters. Retries up to 2 times if the LLM produces malformed output.

pub async fn generate_thread_with_structure( &self, topic: &str, structure: Option<ThreadStructure>, ) -> Result<Vec<String>, LlmError>

Generate a thread using a specific structure for varied content.