rewrite cli parsing; need to implement format-citekeys cli parsing

3 files changed, 124 insertions, 63 deletions

diff --git a/src/bibiman/citekeys.rs b/src/bibiman/citekeys.rs
index b389da2..b7995ac 100644
--- a/src/bibiman/citekeys.rs
+++ b/src/bibiman/citekeys.rs
@@ -16,7 +16,7 @@
 /////
 
 use std::{
-    fs::File,
+    fs::OpenOptions,
     io::Write,
     path::{Path, PathBuf},
 };
@@ -38,7 +38,9 @@ pub enum CitekeyCase {
 
 #[derive(Debug, Default, Clone)]
 pub(crate) struct CitekeyFormatting {
-    bibfile_path: PathBuf,
+    /// bibfile to replace keys at. The optional fields defines a differing
+    /// output file to write to, otherwise original file will be overwritten.
+    bibfile_path: (PathBuf, Option<PathBuf>),
     bib_entries: Bibliography,
     fields: Vec<String>,
     case: Option<CitekeyCase>,
@@ -54,6 +56,7 @@ impl CitekeyFormatting {
     pub fn new<P: AsRef<Path>>(
         cfg: &BibiConfig,
         path: P,
+        target: Option<P>,
         bib_entries: Bibliography,
     ) -> color_eyre::Result<Self> {
         let fields = cfg
@@ -68,7 +71,10 @@ impl CitekeyFormatting {
             ));
         }
         Ok(Self {
-            bibfile_path: path.as_ref().to_path_buf(),
+            bibfile_path: (
+                path.as_ref().to_path_buf(),
+                target.map(|p| p.as_ref().to_path_buf()),
+            ),
             bib_entries,
             fields,
             case: cfg.citekey_formatter.case.clone(),
@@ -77,9 +83,9 @@ impl CitekeyFormatting {
     }
 
     /// Process the actual formatting. The citekey of every entry will be updated.
-    pub fn do_formatting(&mut self) {
+    pub fn do_formatting(&mut self) -> &mut Self {
         let mut old_new_keys: Vec<(String, String)> = Vec::new();
-        for entry in self.bib_entries.iter_mut() {
+        for entry in self.bib_entries.iter() {
             old_new_keys.push((
                 entry.key.clone(),
                 build_citekey(entry, &self.fields, self.case.as_ref()),
@@ -87,16 +93,47 @@ impl CitekeyFormatting {
         }
 
         self.old_new_keys_map = old_new_keys;
+
+        self
     }
 
     /// Write entries with updated citekeys to bibfile
     pub fn update_file(&self) -> color_eyre::Result<()> {
-        let mut file = File::open(&self.bibfile_path)?;
+        let source_file = self.bibfile_path.0.as_path();
+        let target_file = if let Some(path) = &self.bibfile_path.1 {
+            path
+        } else {
+            source_file
+        };
+        let mut content = std::fs::read_to_string(source_file)?;
 
-        file.write_all(self.bib_entries.to_biblatex_string().as_bytes())?;
+        for (old_key, new_key) in self.old_new_keys_map.iter() {
+            content = content.replace(old_key, new_key);
+        }
+
+        let mut new_file = OpenOptions::new()
+            .truncate(true)
+            .write(true)
+            .create(true)
+            .open(target_file)?;
+
+        new_file.write_all(content.as_bytes())?;
 
         Ok(())
     }
+
+    /// Sort the vector containing old/new citekey pairs by the length of the latter.
+    /// That will prevent the replacement longer key parts that equal a full shorter
+    /// key.
+    ///
+    /// You are **very encouraged** to call this method before `update_file()` to
+    /// prevent replacing citekeys partly which afterwards wont match the pattern
+    /// anymore.
+    pub fn rev_sort_new_keys_by_len(&mut self) -> &mut Self {
+        self.old_new_keys_map
+            .sort_by(|a, b| b.1.len().cmp(&a.1.len()));
+        self
+    }
 }
 
 /// Build the citekey from the patterns defined in the config file
@@ -272,6 +309,18 @@ mod tests {
     #[test]
     fn format_citekey_test() {
         let src = r"
+        @article{bos_latex_metadata_and_publishing_workflows_2023,
+            title         = {{LaTeX}, metadata, and publishing workflows},
+            author        = {Bos, Joppe W. and {McCurley}, Kevin S.},
+            year          = {2023},
+            month         = apr,
+            journal       = {arXiv},
+            number        = {{arXiv}:2301.08277},
+            doi           = {10.48550/arXiv.2301.08277},
+            url           = {http://arxiv.org/abs/2301.08277},
+            urldate       = {2023-08-22},
+            note          = {type: article},
+        }
         @book{bhambra_colonialism_social_theory_2021,
             title         = {Colonialism and \textbf{Modern Social Theory}},
             author        = {Bhambra, Gurminder K. and Holmwood, John},
@@ -282,7 +331,7 @@ mod tests {
         ";
         let bibliography = Bibliography::parse(src).unwrap();
         let mut formatting_struct = CitekeyFormatting {
-            bibfile_path: PathBuf::new(),
+            bibfile_path: (PathBuf::new(), None),
             bib_entries: bibliography,
             fields: vec![
                 "entrytype;;;;:".into(),
@@ -294,29 +343,35 @@ mod tests {
             case: None,
             old_new_keys_map: Vec::new(),
         };
-        formatting_struct.do_formatting();
-        let keys = formatting_struct.bib_entries.keys().collect_vec();
+        let _ = formatting_struct.do_formatting();
+        assert_eq!(
+            formatting_struct.old_new_keys_map.get(0).unwrap().1,
+            "article:Bos-McCurley_LaT_met_and_pub_Empt_2023"
+        );
         assert_eq!(
-            keys[0],
+            formatting_struct.old_new_keys_map.get(1).unwrap().1,
             "book:Bhambra-Holmwood_Col_and_Mod_Soc_Camb:and:Medf_2021"
         );
         formatting_struct.case = Some(CitekeyCase::Lower);
-        formatting_struct.do_formatting();
-        let keys = formatting_struct.bib_entries.keys().collect_vec();
+        let _ = formatting_struct.do_formatting().rev_sort_new_keys_by_len();
+        // now the longer citekey is processed first and its in lowercase!
         assert_eq!(
-            keys[0],
+            formatting_struct.old_new_keys_map.get(0).unwrap().1,
             "book:bhambra-holmwood_col_and_mod_soc_camb:and:medf_2021"
         );
-        // let bib_string = formatting_struct.bib_entries.to_biblatex_string();
-        // let new_entry = r"
-        // @book{book:Bhambra-Holmwood_Col_and_Mod_Soc_Camb:and:Medf_2021,
-        //     title         = {Colonialism and \textbf{Modern Social Theory}},
-        //     author        = {Bhambra, Gurminder K. and Holmwood, John},
-        //     location      = {Cambridge and Medford},
-        //     publisher     = {Polity Press},
-        //     date          = {2021},
-        // }
-        // ";
-        // assert_eq!(new_entry, bib_string);
+    }
+
+    #[test]
+    fn sorting_appended_citekeys() {
+        let mut keys: Vec<(String, String)> = vec![
+            ("smith2000".into(), "smith_book_2000".into()),
+            ("smith2000a".into(), "smith_book_2000a".into()),
+            ("smith2000ab".into(), "smith_book_2000ab".into()),
+        ];
+        keys.sort_by(|a, b| b.1.len().cmp(&a.1.len()));
+        let mut keys = keys.iter();
+        assert_eq!(keys.next().unwrap().1, "smith_book_2000ab");
+        assert_eq!(keys.next().unwrap().1, "smith_book_2000a");
+        assert_eq!(keys.next().unwrap().1, "smith_book_2000");
     }
 }
diff --git a/src/cliargs.rs b/src/cliargs.rs
index 082ecda..3b12fc3 100644
--- a/src/cliargs.rs
+++ b/src/cliargs.rs
@@ -18,20 +18,19 @@
 use color_eyre::eyre::Result;
 use dirs::{config_dir, home_dir};
 use lexopt::prelude::*;
+use owo_colors::OwoColorize;
 use owo_colors::colors::css::LightGreen;
 use owo_colors::colors::*;
-use owo_colors::OwoColorize;
 use std::env;
 use std::path::PathBuf;
 use walkdir::WalkDir;
 
 use crate::app;
+use crate::config::BibiConfig;
 
 // struct for CLIArgs
 #[derive(Debug, Default, Clone)]
 pub struct CLIArgs {
-    pub helparg: bool,
-    pub versionarg: bool,
     pub pos_args: Vec<PathBuf>,
     pub cfg_path: Option<PathBuf>,
     pub light_theme: bool,
@@ -39,7 +38,7 @@ pub struct CLIArgs {
 }
 
 impl CLIArgs {
-    pub fn parse_args() -> Result<CLIArgs, lexopt::Error> {
+    pub fn parse_args() -> color_eyre::Result<(CLIArgs, BibiConfig)> {
         let mut args = CLIArgs::default();
         let mut parser = lexopt::Parser::from_env();
 
@@ -52,22 +51,57 @@ impl CLIArgs {
             None
         };
 
+        // if parser
+        //     .raw_args()
+        //     .is_ok_and(|mut arg| arg.next_if(|a| a == "format-citekeys").is_some())
+        // {
+        //     todo!("Format citekeys options");
+        // }
+
         while let Some(arg) = parser.next()? {
             match arg {
-                Short('h') | Long("help") => args.helparg = true,
-                Short('v') | Long("version") => args.versionarg = true,
+                Short('h') | Long("help") => {
+                    println!("{}", help_func());
+                    std::process::exit(0);
+                }
+                Short('v') | Long("version") => {
+                    println!("{}", version_func());
+                    std::process::exit(0);
+                }
                 Short('c') | Long("config-file") => args.cfg_path = Some(parser.value()?.parse()?),
                 Long("light-terminal") => args.light_theme = true,
                 Long("pdf-path") => {
                     args.pdf_path = Some(parser.value()?.parse()?);
                 }
                 // Value(pos_arg) => parse_files(&mut args, pos_arg),
-                Value(pos_arg) => args.pos_args.push(pos_arg.into()),
-                _ => return Err(arg.unexpected()),
+                Value(pos_arg) => {
+                    if args.pos_args.is_empty() && pos_arg == "format-citekeys" {
+                        todo!("Write format citekeys function");
+                    } else {
+                        args.pos_args.push(parser.value()?.into());
+                    }
+                }
+                _ => return Err(arg.unexpected().into()),
             }
         }
 
-        Ok(args)
+        if args
+            .cfg_path
+            .as_ref()
+            .is_some_and(|f| !f.try_exists().unwrap() || !f.is_file())
+        {
+            BibiConfig::create_default_config(&args);
+        }
+
+        let mut cfg = if args.cfg_path.is_some() {
+            BibiConfig::parse_config(&args)?
+        } else {
+            BibiConfig::new(&args)
+        };
+
+        cfg.cli_overwrite(&args);
+
+        Ok((args, cfg))
     }
 }
 
diff --git a/src/main.rs b/src/main.rs
index c956d7c..58805d5 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -31,35 +31,7 @@ pub mod tui;
 #[tokio::main]
 async fn main() -> Result<()> {
     // Parse CLI arguments
-    let mut parsed_args = CLIArgs::parse_args()?;
-
-    // Print help if -h/--help flag is passed and exit
-    if parsed_args.helparg {
-        println!("{}", cliargs::help_func());
-        std::process::exit(0);
-    }
-
-    // Print version if -v/--version flag is passed and exit
-    if parsed_args.versionarg {
-        println!("{}", cliargs::version_func());
-        std::process::exit(0);
-    }
-
-    if parsed_args
-        .cfg_path
-        .as_ref()
-        .is_some_and(|f| !f.try_exists().unwrap() || !f.is_file())
-    {
-        BibiConfig::create_default_config(&parsed_args);
-    }
-
-    let mut cfg = if parsed_args.cfg_path.is_some() {
-        BibiConfig::parse_config(&parsed_args)?
-    } else {
-        BibiConfig::new(&parsed_args)
-    };
-
-    cfg.cli_overwrite(&parsed_args);
+    let (mut parsed_args, mut cfg) = CLIArgs::parse_args()?;
 
     init_error_hooks()?;