feat: xml

- add new feature: xml parser

Signed-off-by: Pakin <pakin.t@forth.co.th>

Cargo.lock

@@ -341,9 +341,9 @@ dependencies = [
 
 [[package]]
 name = "hashbrown"
-version = "0.16.1"
+version = "0.17.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "841d1cc9bed7f9236f321df977030373f4a4163ae1a7dbfe1a51a2c1a51d9100"
+checksum = "4f467dd6dccf739c208452f8014c75c18bb8301b050ad1cfb27153803edb0f51"
 
 [[package]]
 name = "hmac"
@@ -482,9 +482,9 @@ dependencies = [
 
 [[package]]
 name = "indexmap"
-version = "2.13.0"
+version = "2.14.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7714e70437a7dc3ac8eb7e6f8df75fd8eb422675fc7678aff7364301092b1017"
+checksum = "d466e9454f08e4a911e14806c24e16fba1b4c121d1ea474396f396069cf949d9"
 dependencies = [
  "equivalent",
  "hashbrown",
@@ -583,7 +583,9 @@ dependencies = [
  "chrono",
  "flate2",
  "git2",
+ "indexmap",
  "log",
+ "quick-xml",
  "rand",
  "rayon",
  "serde",
@@ -758,6 +760,15 @@ dependencies = [
  "unicode-ident",
 ]
 
+[[package]]
+name = "quick-xml"
+version = "0.39.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "958f21e8e7ceb5a1aa7fa87fab28e7c75976e0bfe7e23ff069e0a260f894067d"
+dependencies = [
+ "memchr",
+]
+
 [[package]]
 name = "quote"
 version = "1.0.43"

Cargo.toml

@@ -7,7 +7,9 @@ edition = "2024"
 chrono = "0.4.41"
 flate2 = "1.1.2"
 git2 = "0.20.2"
+indexmap = "2.14.0"
 log = "0.4.27"
+quick-xml = "0.39.2"
 rand = "0.9.2"
 rayon = "1.10.0"
 serde = { version = "1.0.219", features = ["derive", "serde_derive"] }

README.md

@@ -26,6 +26,7 @@ let recipe_dir = cfg.get("RECIPE_DIR").unwrap();
 ---
 
 ### Get recipe from specific country (latest)
+
 ```rust
 ...
 let latest_versions = grep_latest_versions(recipe_dir).unwrap();
@@ -69,7 +70,6 @@ import::generate_recipe_sheet_table("mys", 626);
 
 ### Notes
 
-
 Simple Snippet Patterns
 
 ```rust
@@ -88,3 +88,224 @@ let pure_mat_id = if curr_rpl_mat_id > 300000 {
     curr_rpl_mat_id
 };
 ```
+
+---
+
+### XML Parser (Experimental)
+
+This will parse xml file and create node structure type `Vec<Node>` where `Node` is from `crate::xml::node`
+
+Xml-related functions
+
+- `parse_xml_to_tree`: this will parse raw xml string into node structure.
+- `print_tree`: printing node structure to xml
+- `generate_nodes_from_xml`: generate node index map from list of `(catalog_name, catalog_path)`
+
+Node functions
+
+- `find_by_child_value`: search expected value from node (`parent_node`) which this value should be in node named `child_name`.
+- `get_child`: get child node from current node's children
+
+Shortcut Macro
+
+- `get_path`: macro for accessing the node inner child by provided key `key1.key2.key3...`
+
+Example of parsing xml layout v3 and generate into `new-layout-v2` format
+
+```rust
+use libtbr::xml::*;
+
+
+// ...
+
+let taobin_dir = cfg
+    .get("TAOBIN_REPO")
+    .expect("Taobin directory path not provided");
+let v3_dir = format!("{taobin_dir}/inter/ltu/xml/multi/v3");
+
+// pre-defined paths configuration in format (catalog_name, catalog_path)
+let v3_catalogs = vec![
+    (
+        "recommend",
+        format!("{ltu_v3_dir}/event/event_v3/active_promotions.lxml"),
+    ),
+    (
+        "coffee",
+        format!("{ltu_v3_dir}/page_catalog_group_coffee.lxml"),
+    ),
+    ("milk", format!("{ltu_v3_dir}/page_catalog_group_milk.lxml")),
+    ("tea", format!("{ltu_v3_dir}/page_catalog_group_tea.lxml")),
+    (
+        "health",
+        format!("{ltu_v3_dir}/page_catalog_group_health.lxml"),
+    ),
+    (
+        "other",
+        format!("{ltu_v3_dir}/page_catalog_group_other.lxml"),
+    ),
+];
+
+// input must be `Vec<(&str, String)>`
+let mut v3_catalog_nodes: IndexMap<String, Vec<Node>> =
+    generate_nodes_from_xml(v3_catalogs)?;
+
+for (_, (catalog_name, catalog_nodes)) in v3_catalog_nodes.iter().enumerate() {
+    if catalog_nodes.len() == 1
+        && let Some(root_node) = catalog_nodes.first()
+    {
+        // get_path is a macro for accessing the node inner child
+        //
+        // usage: get_path!(root_node, key1.key2.key3...);
+        //
+        let current_menus_result: Option<&Node> = get_path!(root_node, ScrollableCatalog.Menus);
+        if let Some(current_menus) = current_menus_result {
+            println!(
+                "Name={},file=page_catalog_group_{}.skt",
+                catalog_name, catalog_name
+            );
+
+            let ccm = current_menus.clone();
+
+            for menu_block in ccm.children.clone() {
+                let mut name_row = String::from("\tname\t");
+                let mut desc_row = String::from("\tdesc\t");
+                let mut img_row = String::from("\timg\t");
+
+                let tag_filter_option = get_path!(menu_block, TagFilter);
+                let idle_image_tag = match get_path!(menu_block, IdleImage) {
+                    Some(img_path) => {
+                        let img_path = img_path.clone().value.unwrap_or("".to_string());
+                        let img_path_spl: Vec<String> = img_path
+                            .trim()
+                            .replace("\"", "")
+                            .split("/")
+                            .map(|x| x.to_string())
+                            .collect();
+                        img_path_spl
+                            .last()
+                            .unwrap()
+                            .replace("[amp]", "&")
+                            .to_string()
+                    }
+                    None => "".to_string(),
+                };
+
+                img_row.push_str(format!("{idle_image_tag}\t-\t-\t-\t\t\t||||||||||||||||||||||||||\t||||||||||||||||||||||||||\t||||||||||||||||||||||||||\t\t\t\t\t\t\t\t-\t-\t-\t-\t-\t").as_str());
+
+                let hot_state_val = match get_path!(menu_block, HotState) {
+                    Some(state) => state
+                        .clone()
+                        .value
+                        .and_then(|x| {
+                            if x.to_string().contains("Disable2") {
+                                return Some("-".to_string());
+                            } else {
+                                return Some(x.replace("$", "").replace(".Button", ""));
+                            }
+                        })
+                        .unwrap()
+                        .trim()
+                        .to_string(),
+                    None => "-".to_string(),
+                };
+                let ice_state_val = match get_path!(menu_block, IceState) {
+                    Some(state) => state
+                        .clone()
+                        .value
+                        .and_then(|x| {
+                            if x.to_string().contains("Disable2") {
+                                return Some("-".to_string());
+                            } else {
+                                return Some(x.replace("$", "").replace(".Button", ""));
+                            }
+                        })
+                        .unwrap()
+                        .trim()
+                        .to_string(),
+                    None => "-".to_string(),
+                };
+                let blend_state_val = match get_path!(menu_block, BlendState) {
+                    Some(state) => state
+                        .clone()
+                        .value
+                        .and_then(|x| {
+                            if x.to_string().contains("Disable2") {
+                                return Some("-".to_string());
+                            } else {
+                                return Some(x.replace("$", "").replace(".Button", ""));
+                            }
+                        })
+                        .unwrap()
+                        .trim()
+                        .to_string(),
+                    None => "-".to_string(),
+                };
+
+                let names = match get_path!(menu_block, Name.LanguageGroup) {
+                    Some(names) => names.clone(),
+                    None => Node::default(),
+                };
+
+                // Description
+                let descs = match get_path!(menu_block, Description.LanguageGroup) {
+                    Some(descs) => descs.clone(),
+                    None => Node::default(),
+                };
+
+                for name in names.children.clone() {
+                    if let Some(value) = name.value {
+                        name_row.push_str(format!("{value}\t").replace("[amp]", "&").as_str());
+                    } else {
+                        name_row.push_str(format!("\t").as_str());
+                    }
+                }
+
+                name_row.push_str(
+                    format!(
+                        "{},-\t{},-\t{},-\t\t\t\t\t\t\t\t-\t-\t-\t-\t{}",
+                        hot_state_val,
+                        ice_state_val,
+                        blend_state_val,
+                        tag_filter_option
+                            .clone()
+                            .unwrap_or(&Node::default())
+                            .value
+                            .clone()
+                            .unwrap_or("-".to_string())
+                            .trim()
+                            .replace("\"", "")
+                    )
+                    .as_str(),
+                );
+
+                for desc in descs.children.clone() {
+                    if let Some(value) = desc.value {
+                        desc_row.push_str(format!("{value}\t").replace("[amp]", "&").as_str());
+                    } else {
+                        desc_row.push_str(format!("\t").as_str());
+                    }
+                }
+
+                desc_row.push_str(
+                    format!(
+                        "||||||||||||||||||||||||||\t||||||||||||||||||||||||||\t||||||||||||||||||||||||||\t\t\t\t\t\t\t\t-\t-\t-\t-\t-\t"
+                    )
+                    .as_str(),
+                );
+
+                //||||||||||||||||||||||||||
+
+                println!("{name_row}");
+                println!("{desc_row}");
+                println!("{img_row}");
+
+                println!("");
+            }
+
+            // name	Americano	อเมริกาโน	Amerikano	Americano			59-01-01-0003,59-21-01-0003	59-01-02-0001,59-21-02-0001	-,-								-	-	Signature	-	CoffeeNoMilk,Recommend
+            // desc	Espresso, Water	กาแฟ และน้ำ	Espresas, vanduo	Espresso, Apă			||||||||||||||||||||||||||	||||||||||||||||||||||||||	||||||||||||||||||||||||||								-	-	-	-	-
+            // img	bn_hot_americano.png	-	bn_hot_america_no.png	bn_hot_america_no.png			posi1	||||||||||||||||||||||||||	||||||||||||||||||||||||||								-	-	-	-	-
+        }
+    }
+}
+```

src/lib.rs

@@ -2,3 +2,4 @@
 pub mod models;
 pub mod previews;
 pub mod recipe_functions;
+pub mod xml;

src/recipe_functions/common.rs

@@ -339,3 +339,12 @@ pub fn grep_latest_versions(dir_path: &str) -> Result<HashMap<String, usize>, st
 
     Ok(vs)
 }
+
+pub fn read_tsv_file(path: &str) -> Result<Vec<String>, Box<dyn std::error::Error>> {
+    let mut file = File::open(path)?;
+    let mut content = String::new();
+
+    file.read_to_string(&mut content)?;
+
+    Ok(content.lines().map(|x| x.to_string()).collect())
+}

src/xml/mod.rs

@@ -0,0 +1 @@
+pub mod node;

src/xml/node.rs

@@ -0,0 +1,158 @@
+use indexmap::IndexMap;
+use quick_xml::events::Event;
+use std::fs::File;
+use std::io::Read;
+
+#[derive(Debug, Clone, Default)]
+pub struct Node {
+    pub name: String,
+    pub children: Vec<Node>,
+    pub value: Option<String>,
+}
+
+pub fn parse_xml_to_tree(xml: &str) -> Vec<Node> {
+    let mut reader = quick_xml::Reader::from_str(xml);
+
+    let mut stack: Vec<Node> = Vec::new();
+    let mut roots: Vec<Node> = Vec::new();
+    let mut buf = Vec::new();
+
+    loop {
+        match reader.read_event_into(&mut buf) {
+            Ok(Event::Start(e)) => {
+                let name = String::from_utf8_lossy(e.name().as_ref()).into_owned();
+                stack.push(Node {
+                    name,
+                    children: Vec::new(),
+                    value: None,
+                });
+            }
+            Ok(Event::End(_)) => {
+                if let Some(finished_node) = stack.pop() {
+                    if let Some(parent) = stack.last_mut() {
+                        parent.children.push(finished_node);
+                    } else {
+                        roots.push(finished_node);
+                    }
+                }
+            }
+            Ok(Event::Text(e)) => {
+                if let Some(current_node) = stack.last_mut() {
+                    let curr_text = String::from_utf8(e.clone().into_inner().to_vec()).unwrap();
+                    // println!("detect text: {curr_text} -- {e:?}");
+                    current_node.value = Some(curr_text);
+                }
+            }
+            Ok(Event::Eof) => break,
+            Err(e) => {
+                println!("error: {e:?}");
+            }
+            _ => {}
+        }
+        buf.clear();
+    }
+    roots
+}
+
+pub fn print_tree(nodes: &[Node], depth: usize) {
+    for node in nodes {
+        let indent = "  ".repeat(depth);
+        match (&node.value, node.children.is_empty()) {
+            (Some(val), true) => {
+                println!("{}<{}>{}</{}>", indent, node.name, val, node.name);
+            }
+            (_, false) => {
+                println!("{}<{}>", indent, node.name);
+                print_tree(&node.children, depth + 1);
+                println!("{}</{}>", indent, node.name);
+            }
+            (None, true) => {
+                println!("{}<{}></{}>", indent, node.name, node.name);
+            }
+        }
+    }
+}
+
+impl Node {
+    pub fn find_by_child_value(
+        &self,
+        parent_name: &str,
+        child_name: &str,
+        target_value: &str,
+    ) -> Vec<Node> {
+        let mut matches = Vec::new();
+
+        if self.name == parent_name {
+            if self.children.iter().any(|c| {
+                c.name == child_name && c.clone().value.is_some_and(|x| x.contains(target_value))
+            }) {
+                matches.push(self.clone());
+            }
+        }
+
+        for child in &self.children {
+            matches.extend(child.find_by_child_value(parent_name, child_name, target_value));
+        }
+        matches
+    }
+
+    pub fn get_child(&self, name: &str) -> Option<&Node> {
+        self.children.iter().find(|c| c.name == name)
+    }
+}
+
+pub fn generate_nodes_from_xml(
+    catalog_map: Vec<(&str, String)>,
+) -> Result<IndexMap<String, Vec<Node>>, Box<dyn std::error::Error>> {
+    let mut result = IndexMap::new();
+
+    for catalog_m in catalog_map {
+        let catalog_name = catalog_m.0;
+        let catalog_path = catalog_m.1.clone();
+
+        let mut file = File::open(catalog_path)?;
+        let mut content = String::new();
+
+        file.read_to_string(&mut content)?;
+
+        // clean comment
+        let mut new_file = String::new();
+        for line in content.lines() {
+            if line.contains(";") && !line.contains(";include") {
+                continue;
+            } else {
+                new_file.push_str(format!("{line}\n").replace("&", "[amp]").as_str());
+            }
+        }
+
+        let node = parse_xml_to_tree(&new_file);
+        result.insert(catalog_name.to_string(), node);
+    }
+
+    Ok(result)
+}
+
+/// get node from path of node vector
+///
+/// Example:
+/// ```
+/// let current_menus_result: Option<&Node> = get_path!(root_node, ScrollableCatalog.Menus);
+///
+/// // Possible results
+///
+/// //Some(Node { name: "Menus", children: [Node { name: "Menu", children: [Node { name: "State",....
+///
+/// //None
+/// ```
+#[macro_export]
+macro_rules! get_path {
+    ($node:expr, $last:ident) => {
+        $node.get_child(stringify!($last))
+    };
+    // recursive case
+        ($node:expr, $next:ident . $($rest:ident).+) => {
+            $node
+                .get_child(stringify!($next))
+                .and_then(|child| get_path!(child, $($rest).+))
+        };
+}