jetzt werden die Sektionen aus einer config.yml rausgezogen.

3 files changed, 71 insertions, 20 deletions

diff --git a/src/config.rs b/src/config.rs
new file mode 100644
index 0000000..fc8ee03
--- /dev/null
+++ b/src/config.rs
@@ -0,0 +1,44 @@
+extern crate yaml_rust;
+use self::yaml_rust::YamlLoader;
+
+use std::collections::HashMap;
+use std::fs::File;
+use std::io::prelude::*;
+
+pub struct SectionConfig {
+    pub pattern: String,
+    pub groups: Vec<String>,
+}
+
+pub struct Config {
+    pub sections: HashMap<String, SectionConfig>,
+}
+
+impl Config {
+    pub fn from_file(p: &str) -> Self {
+        let mut f = File::open(p).unwrap();
+        let mut buf = String::new();
+        f.read_to_string(&mut buf).unwrap();
+        let docs = YamlLoader::load_from_str(&buf).unwrap();
+
+        let doc = &docs[0];
+
+        println!("{:?}", doc);
+
+        let mut sections: HashMap<String, SectionConfig> = HashMap::new();
+        for (name, entry) in doc["sections"].as_hash().unwrap() {
+            sections.insert(name.as_str().unwrap().into(),
+                            SectionConfig {
+                                pattern: entry["pattern"].as_str().unwrap().into(),
+                                groups: entry["groups"]
+                                    .as_vec()
+                                    .unwrap()
+                                    .into_iter()
+                                    .map(|v| v.as_str().unwrap().into())
+                                    .collect(),
+                            });
+        }
+
+        Config { sections: sections }
+    }
+}
diff --git a/src/main.rs b/src/main.rs
index 97b3094..7d06fe7 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -1,4 +1,5 @@
 extern crate regex;
+extern crate yaml_rust;
 
 extern crate walkdir;
 use walkdir::WalkDir;
@@ -12,13 +13,19 @@ use std::collections::HashMap;
 mod pre_process;
 use pre_process::Section;
 
+mod config;
+use config::Config;
+
 fn main() {
     let raw_files = env::var("RAW_FILES").unwrap_or("characters".into());
     let base_path = Path::new(&raw_files);
 
+    let cfg = Config::from_file("config.yml");
+
     let mut sections: Vec<Section> = vec![];
-        sections.push(Section::new("name", r#"(?is)Romaji Name.*?<TD>(.*?)\s?</TD>.*?Japanese Name.*?<TD>(.*?)\s?</TD>"#, vec!["romaji".into(), "japanese".into()]));
-    sections.push(Section::new("image", r#"(?is)<H3 id="section99">.*<img src="(.*?)" alt=.*?></a><p><a href="(.*?)">View Full Size Image"#, vec!["thumb".into(), "full".into()]));
+    for (name, sec) in &cfg.sections {
+        sections.push(Section::new(&name, &sec.pattern, sec.groups.clone()));
+    }
 
     for entry in WalkDir::new(base_path).min_depth(1).into_iter().filter_map(|e| e.ok()) {
         let mut f = File::open(entry.path()).expect("could not open file");
@@ -32,11 +39,11 @@ fn main() {
 
         pre_process::split_sections(&buf, &mut sections);
 
-                let mut char: HashMap<String, HashMap<String, String>> = HashMap::new();
-                for s in &sections {
-                        char.insert(s.name.clone(), s.data.clone());
-                }
+        let mut char: HashMap<String, HashMap<String, String>> = HashMap::new();
+        for s in &sections {
+            char.insert(s.name.clone(), s.data.clone());
+        }
 
-                println!("{:?}", char);
+        println!("{:?}", char);
     }
 }
diff --git a/src/pre_process.rs b/src/pre_process.rs
index 0d8c6be..a4d6c14 100644
--- a/src/pre_process.rs
+++ b/src/pre_process.rs
@@ -24,14 +24,14 @@ pub struct Section {
 }
 
 impl Section {
-        pub fn new(name: &str, re: &str, groups: Vec<String>) -> Self {
-                Section {
-                        name: name.into(),
-                        re: Regex::new(re).unwrap(),
-                        keys: groups,
-                        data: HashMap::new(),
-                }
-        }
+    pub fn new(name: &str, re: &str, groups: Vec<String>) -> Self {
+        Section {
+            name: name.into(),
+            re: Regex::new(re).unwrap(),
+            keys: groups,
+            data: HashMap::new(),
+        }
+    }
 }
 
 pub fn split_sections(d: &str, s: &mut Vec<Section>) {
@@ -39,11 +39,11 @@ pub fn split_sections(d: &str, s: &mut Vec<Section>) {
         for m in section.re.captures_iter(d) {
             assert!(m.len() >= section.keys.len() + 1);
 
-                        let mut idx = 0;
-                        for key in &section.keys {
-                                section.data.insert(key.clone(), m.at(idx + 1).unwrap().into());
-                                idx += 1;
-                        }
+            let mut idx = 0;
+            for key in &section.keys {
+                section.data.insert(key.clone(), m.at(idx + 1).unwrap().into());
+                idx += 1;
+            }
         }
     }
 }