edgeparse 0.1.0__tar.gz → 0.1.1__tar.gz

{edgeparse-0.1.0 → edgeparse-0.1.1}/Cargo.lock

@@ -608,7 +608,7 @@ dependencies = [
 
 [[package]]
 name = "edgeparse-cli"
-version = "0.1.0"
+version = "0.1.1"
 dependencies = [
  "anyhow",
  "clap",
@@ -622,7 +622,7 @@ dependencies = [
 
 [[package]]
 name = "edgeparse-core"
-version = "0.1.0"
+version = "0.1.1"
 dependencies = [
  "anyhow",
  "base64",
@@ -646,7 +646,7 @@ dependencies = [
 
 [[package]]
 name = "edgeparse-node"
-version = "0.1.0"
+version = "0.1.1"
 dependencies = [
  "edgeparse-core",
  "napi",
@@ -658,7 +658,7 @@ dependencies = [
 
 [[package]]
 name = "edgeparse-python"
-version = "0.1.0"
+version = "0.1.1"
 dependencies = [
  "edgeparse-core",
  "pyo3",

{edgeparse-0.1.0 → edgeparse-0.1.1}/Cargo.toml

@@ -3,11 +3,11 @@ resolver = "2"
 members = ["crates/edgeparse-core", "crates/pdf-cos", "crates/edgeparse-python"]
 
 [workspace.package]
-version = "0.1.0"
+version = "0.1.1"
 edition = "2021"
 rust-version = "1.85"
 license = "Apache-2.0"
-repository = "https://github.com/opendataloader-project/edgeparse"
+repository = "https://github.com/raphaelmansuy/edgeparse"
 description = "EdgeParse — High-performance PDF-to-structured-data extraction engine"
 
 [workspace.dependencies]

{edgeparse-0.1.0 → edgeparse-0.1.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: edgeparse
-Version: 0.1.0
+Version: 0.1.1
 Classifier: Development Status :: 4 - Beta
 Classifier: Intended Audience :: Developers
 Classifier: License :: OSI Approved :: Apache Software License
@@ -17,9 +17,9 @@ Summary: High-performance PDF-to-structured-data extraction — Rust engine, Pyt
 Keywords: pdf,extraction,markdown,rag,llm
 Requires-Python: >=3.9
 Description-Content-Type: text/markdown; charset=UTF-8; variant=GFM
-Project-URL: Documentation, https://opendataloader-project.github.io/edgeparse
-Project-URL: Homepage, https://github.com/opendataloader-project/edgeparse
-Project-URL: Repository, https://github.com/opendataloader-project/edgeparse
+Project-URL: Documentation, https://github.com/raphaelmansuy/edgeparse/tree/main/docs
+Project-URL: Homepage, https://github.com/raphaelmansuy/edgeparse
+Project-URL: Repository, https://github.com/raphaelmansuy/edgeparse
 
 # edgeparse
 
@@ -68,5 +68,5 @@ edgeparse /path/to/dir/ --output-dir out/  # batch convert
 
 - [GitHub](https://github.com/raphaelmansuy/edgeparse)
 - [crates.io (Rust)](https://crates.io/crates/edgeparse-core)
-- [npm (@edgeparse/pdf)](https://www.npmjs.com/package/@edgeparse/pdf)
+- [npm (edgeparse)](https://www.npmjs.com/package/edgeparse)
 

{edgeparse-0.1.0 → edgeparse-0.1.1}/README.md

@@ -45,4 +45,4 @@ edgeparse /path/to/dir/ --output-dir out/  # batch convert
 
 - [GitHub](https://github.com/raphaelmansuy/edgeparse)
 - [crates.io (Rust)](https://crates.io/crates/edgeparse-core)
-- [npm (@edgeparse/pdf)](https://www.npmjs.com/package/@edgeparse/pdf)
+- [npm (edgeparse)](https://www.npmjs.com/package/edgeparse)

{edgeparse-0.1.0 → edgeparse-0.1.1}/crates/edgeparse-core/src/api/config_loader.rs

@@ -39,7 +39,10 @@ pub fn config_to_json(config: &ProcessingConfig) -> Result<String, EdgePdfError>
 ///
 /// This works by serializing both to JSON, merging the JSON objects, and
 /// deserializing back. Fields in `overlay` that are `null` are skipped.
-pub fn merge_configs(base: &ProcessingConfig, overlay_json: &str) -> Result<ProcessingConfig, EdgePdfError> {
+pub fn merge_configs(
+    base: &ProcessingConfig,
+    overlay_json: &str,
+) -> Result<ProcessingConfig, EdgePdfError> {
     let base_json = serde_json::to_value(base)
         .map_err(|e| EdgePdfError::OutputError(format!("config serialization error: {}", e)))?;
     let overlay_val: serde_json::Value = serde_json::from_str(overlay_json)

edgeparse-0.1.1/crates/edgeparse-core/src/hybrid.rs

@@ -0,0 +1,4 @@
+//! Hybrid extraction mode — combines rule-based pipeline with LLM-assisted
+//! post-processing for higher-accuracy structured output.
+//!
+//! This module is only compiled when the `hybrid` feature flag is enabled.

{edgeparse-0.1.0 → edgeparse-0.1.1}/crates/edgeparse-core/src/lib.rs

@@ -54,7 +54,10 @@ pub fn convert(
     for (&page_num, &page_id) in &pages_map {
         let page_chunks = extract_page_chunks(&raw_doc.document, page_num, page_id)?;
         let mut recovered_tables = Vec::new();
-        if let Some(page_info) = page_info_list.iter().find(|info| info.page_number == page_num) {
+        if let Some(page_info) = page_info_list
+            .iter()
+            .find(|info| info.page_number == page_num)
+        {
             recovered_tables = recover_raster_table_borders(
                 input_path,
                 &page_info.crop_box,
@@ -87,16 +90,19 @@ pub fn convert(
                 .into_iter()
                 .map(ContentElement::LineArt),
         );
-        elements.extend(recovered_tables.into_iter().map(ContentElement::TableBorder));
+        elements.extend(
+            recovered_tables
+                .into_iter()
+                .map(ContentElement::TableBorder),
+        );
 
         page_contents.push(elements);
     }
 
     // Run the processing pipeline
     let mcid_map = build_mcid_map(&raw_doc.document);
-    let mut pipeline_state =
-        PipelineState::with_mcid_map(page_contents, config.clone(), mcid_map)
-            .with_page_info(page_info_list);
+    let mut pipeline_state = PipelineState::with_mcid_map(page_contents, config.clone(), mcid_map)
+        .with_page_info(page_info_list);
     run_pipeline(&mut pipeline_state)?;
 
     // Build the output document
@@ -163,7 +169,10 @@ impl From<lopdf::Error> for EdgePdfError {
 #[cfg(test)]
 mod tests {
     use super::*;
-    use lopdf::{dictionary, content::{Content, Operation}, Object, Stream};
+    use lopdf::{
+        content::{Content, Operation},
+        dictionary, Object, Stream,
+    };
     use std::io::Write;
 
     /// Create a synthetic PDF file for integration testing.
@@ -238,7 +247,10 @@ mod tests {
 
         let doc = result.unwrap();
         assert_eq!(doc.number_of_pages, 1);
-        assert!(!doc.kids.is_empty(), "Expected content elements in document");
+        assert!(
+            !doc.kids.is_empty(),
+            "Expected content elements in document"
+        );
 
         // Check that we extracted content (may be TextChunks, TextLines, or TextBlocks after pipeline)
         let mut all_text = String::new();
@@ -268,11 +280,18 @@ mod tests {
             }
         }
 
-        assert!(all_text.contains("Hello"), "Expected 'Hello' in extracted text, got: {}", all_text);
-        assert!(all_text.contains("Second"), "Expected 'Second' in extracted text, got: {}", all_text);
+        assert!(
+            all_text.contains("Hello"),
+            "Expected 'Hello' in extracted text, got: {}",
+            all_text
+        );
+        assert!(
+            all_text.contains("Second"),
+            "Expected 'Second' in extracted text, got: {}",
+            all_text
+        );
 
         // Cleanup
         let _ = std::fs::remove_file(&pdf_path);
     }
-
 }

{edgeparse-0.1.0 → edgeparse-0.1.1}/crates/edgeparse-core/src/models/bbox.rs

@@ -46,13 +46,7 @@ impl PartialEq for BoundingBox {
 
 impl BoundingBox {
     /// Create a new BoundingBox.
-    pub fn new(
-        page: Option<u32>,
-        left_x: f64,
-        bottom_y: f64,
-        right_x: f64,
-        top_y: f64,
-    ) -> Self {
+    pub fn new(page: Option<u32>, left_x: f64, bottom_y: f64, right_x: f64, top_y: f64) -> Self {
         Self {
             page_number: page,
             last_page_number: page,

{edgeparse-0.1.0 → edgeparse-0.1.1}/crates/edgeparse-core/src/models/text.rs

@@ -147,8 +147,7 @@ impl TextLine {
         };
 
         let mut result = String::with_capacity(
-            real_chunks.iter().map(|c| c.value.len()).sum::<usize>()
-                + real_chunks.len(),
+            real_chunks.iter().map(|c| c.value.len()).sum::<usize>() + real_chunks.len(),
         );
         result.push_str(&real_chunks[0].value);
 
@@ -306,9 +305,9 @@ impl TextBlock {
                 if result.ends_with('-') {
                     // Check it's a real hyphenation (lowercase letter before hyphen)
                     let before_hyphen = result[..result.len() - 1].chars().last();
-                    if before_hyphen.map_or(false, |c| c.is_alphabetic()) {
+                    if before_hyphen.is_some_and(|c| c.is_alphabetic()) {
                         result.pop(); // Remove the hyphen
-                        // Don't add a space — the word continues
+                                      // Don't add a space — the word continues
                     } else {
                         result.push(' ');
                     }

{edgeparse-0.1.0 → edgeparse-0.1.1}/crates/edgeparse-core/src/output/docx.rs

@@ -116,9 +116,7 @@ fn token_rows_text(rows: &[TableTokenRow]) -> String {
 }
 
 fn make_paragraph(text: &str) -> String {
-    format!(
-        "<w:p><w:r><w:t xml:space=\"preserve\">{text}</w:t></w:r></w:p>\n"
-    )
+    format!("<w:p><w:r><w:t xml:space=\"preserve\">{text}</w:t></w:r></w:p>\n")
 }
 
 fn make_heading(text: &str, level: u32) -> String {

{edgeparse-0.1.0 → edgeparse-0.1.1}/crates/edgeparse-core/src/output/html.rs

@@ -62,7 +62,11 @@ fn render_element(out: &mut String, element: &ContentElement) {
             for item in &list.list_items {
                 let label = token_rows_text(&item.label.content);
                 let body = token_rows_text(&item.body.content);
-                out.push_str(&format!("<li>{}{}</li>\n", html_escape(label.trim()), html_escape(body.trim())));
+                out.push_str(&format!(
+                    "<li>{}{}</li>\n",
+                    html_escape(label.trim()),
+                    html_escape(body.trim())
+                ));
             }
             out.push_str("</ul>\n");
         }

{edgeparse-0.1.0 → edgeparse-0.1.1}/crates/edgeparse-core/src/output/json.rs

@@ -7,7 +7,10 @@ use crate::EdgePdfError;
 ///
 /// # Errors
 /// Returns `EdgePdfError::OutputError` if serialization fails.
-pub fn write_json<W: std::io::Write>(doc: &PdfDocument, writer: &mut W) -> Result<(), EdgePdfError> {
+pub fn write_json<W: std::io::Write>(
+    doc: &PdfDocument,
+    writer: &mut W,
+) -> Result<(), EdgePdfError> {
     serde_json::to_writer_pretty(writer, doc)
         .map_err(|e| EdgePdfError::OutputError(format!("JSON serialization failed: {}", e)))
 }

{edgeparse-0.1.0 → edgeparse-0.1.1}/crates/edgeparse-core/src/output/legacy_json.rs

@@ -9,9 +9,8 @@
 
 use std::cell::Cell;
 
-use serde_json::{Map, Value, json};
+use serde_json::{json, Map, Value};
 
-use crate::EdgePdfError;
 use crate::models::bbox::BoundingBox;
 use crate::models::chunks::ImageChunk;
 use crate::models::content::ContentElement;
@@ -23,13 +22,14 @@ use crate::models::semantic::{
     SemanticNumberHeading, SemanticParagraph, SemanticTable, SemanticTextNode,
 };
 use crate::models::table::{TableBorder, TableBorderCell, TableBorderRow, TableToken};
+use crate::EdgePdfError;
 
 // ---------------------------------------------------------------------------
 // Thread-local ID counter
 // ---------------------------------------------------------------------------
 
 thread_local! {
-    static NEXT_ID: Cell<u64> = Cell::new(1);
+    static NEXT_ID: Cell<u64> = const { Cell::new(1) };
 }
 
 fn next_id() -> u64 {
@@ -76,10 +76,7 @@ fn legacy_float_str(v: f64) -> String {
 fn text_color_string(color: &Option<Vec<f64>>) -> String {
     match color {
         Some(components) if !components.is_empty() => {
-            let parts: Vec<String> = components
-                .iter()
-                .map(|v| legacy_float_str(*v))
-                .collect();
+            let parts: Vec<String> = components.iter().map(|v| legacy_float_str(*v)).collect();
             format!("[{}]", parts.join(", "))
         }
         _ => String::new(),
@@ -144,7 +141,7 @@ fn text_node_style(node: &SemanticTextNode) -> (String, f64, String) {
             .unwrap_or("")
             .to_string(),
         font_size.unwrap_or(0.0),
-        text_color_string(&text_color),
+        text_color_string(text_color),
     )
 }
 
@@ -152,11 +149,7 @@ fn text_node_style(node: &SemanticTextNode) -> (String, f64, String) {
 /// First tries semantic contents (preferred), then falls back to raw body tokens.
 fn list_item_text(item: &ListItem) -> String {
     if !item.contents.is_empty() {
-        let parts: Vec<String> = item
-            .contents
-            .iter()
-            .filter_map(element_text_str)
-            .collect();
+        let parts: Vec<String> = item.contents.iter().filter_map(element_text_str).collect();
         if !parts.is_empty() {
             return parts.join(" ").trim().to_string();
         }
@@ -187,15 +180,27 @@ fn element_text_str(el: &ContentElement) -> Option<String> {
         // Raw text elements from list_detector and early pipeline stages
         ContentElement::TextLine(l) => {
             let s = l.value().trim().to_string();
-            if s.is_empty() { None } else { Some(s) }
+            if s.is_empty() {
+                None
+            } else {
+                Some(s)
+            }
         }
         ContentElement::TextBlock(b) => {
             let s = b.value().trim().to_string();
-            if s.is_empty() { None } else { Some(s) }
+            if s.is_empty() {
+                None
+            } else {
+                Some(s)
+            }
         }
         ContentElement::TextChunk(c) => {
             let s = c.value.trim().to_string();
-            if s.is_empty() { None } else { Some(s) }
+            if s.is_empty() {
+                None
+            } else {
+                Some(s)
+            }
         }
         _ => None,
     }
@@ -210,12 +215,7 @@ fn list_item_style(item: &ListItem) -> (String, f64, String) {
         }
     }
     // Try body tokens first chunk
-    let first_chunk = item
-        .body
-        .content
-        .iter()
-        .flat_map(|row| row.iter())
-        .next();
+    let first_chunk = item.body.content.iter().flat_map(|row| row.iter()).next();
     if let Some(token) = first_chunk {
         let color_str = format_font_color(&token.base.font_color);
         return (
@@ -245,7 +245,9 @@ fn element_style(el: &ContentElement) -> Option<(String, f64, String)> {
             }
         }
         ContentElement::TextBlock(b) => {
-            if let Some(chunk) = b.text_lines.iter()
+            if let Some(chunk) = b
+                .text_lines
+                .iter()
                 .flat_map(|l| l.text_chunks.iter())
                 .find(|c| !c.value.trim().is_empty())
             {
@@ -497,10 +499,7 @@ fn list_to_legacy(list: &PDFList, stem: &str, img_idx: &mut u64) -> Value {
     obj.insert("level".into(), json!(level));
     obj.insert("page number".into(), json!(page_num(&list.bbox)));
     obj.insert("bounding box".into(), bbox_array(&list.bbox));
-    obj.insert(
-        "numbering style".into(),
-        json!(numbering),
-    );
+    obj.insert("numbering style".into(), json!(numbering));
     obj.insert("number of list items".into(), json!(num_items));
     obj.insert("next list id".into(), json!(next_list_id_val));
     obj.insert("previous list id".into(), json!(prev_list_id_val));
@@ -672,8 +671,12 @@ fn elements_to_legacy(el: &ContentElement, stem: &str, img_idx: &mut u64) -> Vec
         ContentElement::Figure(fig) => figure_to_legacy(fig, stem, img_idx),
         ContentElement::Image(img) => vec![image_to_legacy(img, stem, img_idx)],
         ContentElement::List(l) => vec![list_to_legacy(l, stem, img_idx)],
-        ContentElement::Table(st) => semantic_table_to_legacy(st, stem, img_idx).into_iter().collect(),
-        ContentElement::TableBorder(tb) => table_border_to_legacy(tb, stem, img_idx).into_iter().collect(),
+        ContentElement::Table(st) => semantic_table_to_legacy(st, stem, img_idx)
+            .into_iter()
+            .collect(),
+        ContentElement::TableBorder(tb) => table_border_to_legacy(tb, stem, img_idx)
+            .into_iter()
+            .collect(),
         // Skip raw low-level elements not exposed in legacy output
         ContentElement::TextChunk(_)
         | ContentElement::TextLine(_)
@@ -705,10 +708,7 @@ pub fn to_legacy_json_value(doc: &PdfDocument, stem: &str) -> Value {
 
     let mut obj = Map::new();
     obj.insert("file name".into(), json!(doc.file_name));
-    obj.insert(
-        "number of pages".into(),
-        json!(doc.number_of_pages),
-    );
+    obj.insert("number of pages".into(), json!(doc.number_of_pages));
     obj.insert(
         "author".into(),
         doc.author.as_deref().map_or(Value::Null, |s| json!(s)),
@@ -719,11 +719,15 @@ pub fn to_legacy_json_value(doc: &PdfDocument, stem: &str) -> Value {
     );
     obj.insert(
         "creation date".into(),
-        doc.creation_date.as_deref().map_or(Value::Null, |s| json!(s)),
+        doc.creation_date
+            .as_deref()
+            .map_or(Value::Null, |s| json!(s)),
     );
     obj.insert(
         "modification date".into(),
-        doc.modification_date.as_deref().map_or(Value::Null, |s| json!(s)),
+        doc.modification_date
+            .as_deref()
+            .map_or(Value::Null, |s| json!(s)),
     );
     obj.insert("kids".into(), json!(kids));
 
@@ -748,18 +752,18 @@ pub fn to_legacy_json_string(doc: &PdfDocument, stem: &str) -> Result<String, Ed
 mod tests {
     use super::*;
     use crate::models::bbox::BoundingBox;
+    use crate::models::enums::SemanticType;
     use crate::models::semantic::{SemanticParagraph, SemanticTextNode};
     use crate::models::text::TextColumn;
-    use crate::models::enums::SemanticType;
 
     fn make_bbox(page: u32, left: f64, bottom: f64, right: f64, top: f64) -> BoundingBox {
         BoundingBox::new(Some(page), left, bottom, right, top)
     }
 
     fn make_text_node(bbox: BoundingBox, text: &str) -> SemanticTextNode {
-        use crate::models::text::{TextBlock, TextLine};
         use crate::models::chunks::TextChunk;
         use crate::models::enums::{PdfLayer, TextFormat, TextType};
+        use crate::models::text::{TextBlock, TextLine};
         let chunk = TextChunk {
             value: text.to_string(),
             bbox: bbox.clone(),
@@ -869,8 +873,14 @@ mod tests {
 
     #[test]
     fn test_text_color_grayscale() {
-        assert_eq!(text_color_string(&Some(vec![0.0, 0.0, 0.0])), "[0.0, 0.0, 0.0]");
-        assert_eq!(text_color_string(&Some(vec![1.0, 1.0, 1.0])), "[1.0, 1.0, 1.0]");
+        assert_eq!(
+            text_color_string(&Some(vec![0.0, 0.0, 0.0])),
+            "[0.0, 0.0, 0.0]"
+        );
+        assert_eq!(
+            text_color_string(&Some(vec![1.0, 1.0, 1.0])),
+            "[1.0, 1.0, 1.0]"
+        );
         assert_eq!(text_color_string(&None), "");
     }