Method extract_docx_content

core/src/document_loader.rs:709–752

Extract content from DOCX files

(file_path: &str)

    /// Extract content from DOCX files
    async fn extract_docx_content(file_path: &str) -> GraphBitResult<String> {
        use std::fs::File;
        use std::io::Read;

        let mut file = File::open(file_path).map_err(|e| {
            GraphBitError::validation("document_loader", format!("Failed to open DOCX file: {e}"))
        })?;

        let mut buffer = Vec::new();
        file.read_to_end(&mut buffer).map_err(|e| {
            GraphBitError::validation("document_loader", format!("Failed to read DOCX file: {e}"))
        })?;

        let docx = docx_rs::read_docx(&buffer).map_err(|e| {
            GraphBitError::validation("document_loader", format!("Failed to parse DOCX file: {e}"))
        })?;

        let mut text_content = String::new();

        // Extract text from document children
        for child in &docx.document.children {
            if let docx_rs::DocumentChild::Paragraph(paragraph) = child {
                for para_child in &paragraph.children {
                    if let docx_rs::ParagraphChild::Run(run_element) = para_child {
                        for run_child in &run_element.children {
                            if let docx_rs::RunChild::Text(text) = run_child {
                                text_content.push_str(&text.text);
                            }
                        }
                    }
                }
                text_content.push('\n');
            }
        }

        if text_content.trim().is_empty() {
            return Err(GraphBitError::validation(
                "document_loader",
                "No text content could be extracted from the DOCX file",
            ));
        }

        Ok(text_content.trim().to_string())
    }

    /// Extract content from Excel (XLSB, XLSX, XLS,etc.) files
    async fn extract_excel_content(file_path: &str) -> GraphBitResult<String> {