From fd454675133a8fb0b5dfd969bf6e4f08452e2bb8 Mon Sep 17 00:00:00 2001
From: Paul-Christian Volkmer <volkmer_p@ukw.de>
Date: Wed, 12 Jun 2024 16:50:29 +0200
Subject: [PATCH] feat: show items not equal within db and xml

---
 src/lkrexport.rs | 19 ++++++++++++++++++-
 src/main.rs      | 46 ++++++++++++++++++++++++++++++++++++++++------
 2 files changed, 58 insertions(+), 7 deletions(-)

diff --git a/src/lkrexport.rs b/src/lkrexport.rs
index 27e3e2b..ee3e25c 100644
--- a/src/lkrexport.rs
+++ b/src/lkrexport.rs
@@ -124,6 +124,11 @@ impl Meldung {
             _ => None,
         }
     }
+
+    pub fn no_linebreak(&self) -> String {
+        let re = Regex::new(r"\n\s*").unwrap();
+        re.replace_all(&self.raw_value, "").trim().to_string()
+    }
 }
 
 pub fn to_database_id(id: &str) -> Option<String> {
@@ -147,7 +152,7 @@ pub fn to_database_id(id: &str) -> Option<String> {
 
 #[cfg(test)]
 mod tests {
-    use crate::lkrexport::LkrExportProtocolFile;
+    use crate::lkrexport::{LkrExportProtocolFile, Meldung};
 
     #[test]
     fn should_read_xml_file_content() {
@@ -204,4 +209,16 @@ mod tests {
             Some("C17.2".to_string())
         );
     }
+
+    #[test]
+    fn should_get_meldung_with_trimmed_margin() {
+        let meldung = Meldung {
+            raw_value: "  <Test>\n  <Test2>TestInhalt 3</Test2>\n</Test>\n".into(),
+        };
+
+        assert_eq!(
+            meldung.no_linebreak(),
+            "<Test><Test2>TestInhalt 3</Test2></Test>".to_string()
+        );
+    }
 }
diff --git a/src/main.rs b/src/main.rs
index b2b7d1e..4c0a91d 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -29,7 +29,7 @@ use itertools::Itertools;
 use crate::cli::{Cli, SubCommand};
 use crate::common::{Check, DiffRecord, Icd10GroupSize};
 use crate::database::DatabaseSource;
-use crate::lkrexport::{to_database_id, LkrExportProtocolFile};
+use crate::lkrexport::{to_database_id, LkrExportProtocolFile, Meldung};
 
 mod cli;
 mod common;
@@ -467,6 +467,11 @@ fn main() -> Result<(), Box<dyn Error>> {
                 .map(|meldung| (meldung.id().unwrap(), meldung))
                 .collect::<HashMap<_, _>>();
 
+            let missing_xml_ids = db_meldungen
+                .keys()
+                .filter(|&key| !xml_meldungen.contains_key(key))
+                .collect_vec();
+
             let _ = term.clear_last_lines(1);
 
             let _ = term.write_line(
@@ -512,11 +517,6 @@ fn main() -> Result<(), Box<dyn Error>> {
                     });
                 }
 
-                let missing_xml_ids = db_meldungen
-                    .keys()
-                    .filter(|&key| !xml_meldungen.contains_key(key))
-                    .collect_vec();
-
                 if !missing_xml_ids.is_empty() {
                     let _ = term.write_line(
                         &style("\nIn der Protokolldatei fehlende Meldungen:")
@@ -558,6 +558,40 @@ fn main() -> Result<(), Box<dyn Error>> {
                     let _ = term.write_line(&item.to_string());
                 });
             }
+
+            let different_content = db_meldungen
+                .iter()
+                .filter(|(id, _)| !missing_xml_ids.contains(id))
+                .filter(|(id, meldung)| {
+                    xml_meldungen
+                        .get(&id.to_string())
+                        .unwrap_or(&Meldung {
+                            raw_value: String::new(),
+                        })
+                        .no_linebreak()
+                        != meldung.no_linebreak()
+                })
+                .map(|(_, meldung)| meldung.id().unwrap_or("?".into()))
+                .collect_vec();
+
+            if !different_content.is_empty() {
+                let _ = term.write_line(
+                    &style(&format!(
+                        "\nFolgende {} Meldungen unterscheiden sich in der Datenbank und der Protokolldatei:",
+                        different_content.len()
+                    ))
+                    .yellow()
+                    .to_string(),
+                );
+
+                different_content.iter().sorted().for_each(|item| {
+                    let _ = term.write_line(&format!(
+                        "{} ({})",
+                        item,
+                        to_database_id(item).unwrap_or("?".into())
+                    ));
+                });
+            }
         }
     }