fix: parsing history dates

codegen/src/collect_history_dates.rs

@@ -5,26 +5,25 @@ use rustypipe::{
     client::RustyPipe,
     param::{Language, LANGUAGES},
 };
-use serde::{Deserialize, Serialize};
 
 use crate::util::{self, DICT_DIR};
 
-type CollectedDates = BTreeMap<Language, HistoryDates>;
+type CollectedDates = BTreeMap<Language, BTreeMap<String, String>>;
 
-#[derive(Debug, Serialize, Deserialize)]
-struct HistoryDates {
-    this_week: String,
-    last_week: String,
-}
+const THIS_WEEK: &str = "this_week";
+const LAST_WEEK: &str = "last_week";
 
-pub async fn collect_dates() {
+pub async fn collect_dates_music() {
     let json_path = path!(*DICT_DIR / "history_date_samples.json");
     let rp = RustyPipe::builder()
-        .storage_dir("/home/thetadev/Documents/Programmieren/Rust/rustypipe")
+        .storage_dir(path!(env!("CARGO_MANIFEST_DIR") / ".."))
         .build()
         .unwrap();
 
-    let mut res: CollectedDates = BTreeMap::new();
+    let mut res: CollectedDates = {
+        let json_file = File::open(&json_path).unwrap();
+        serde_json::from_reader(BufReader::new(json_file)).unwrap()
+    };
 
     for lang in LANGUAGES {
         println!("{lang}");
@@ -34,14 +33,56 @@ pub async fn collect_dates() {
         }
 
         // The indexes have to be adapted before running
-        let d = HistoryDates {
-            this_week: history.items[0].playback_date_txt.clone().unwrap(),
-            last_week: history.items[18].playback_date_txt.clone().unwrap(),
-        };
-        res.insert(lang, d);
+        let entry = res.entry(lang).or_default();
+        entry.insert(
+            THIS_WEEK.to_owned(),
+            history.items[0].playback_date_txt.clone().unwrap(),
+        );
+        entry.insert(
+            LAST_WEEK.to_owned(),
+            history.items[18].playback_date_txt.clone().unwrap(),
+        );
     }
 
-    let file = File::create(json_path).unwrap();
+    let file = File::create(&json_path).unwrap();
+    serde_json::to_writer_pretty(file, &res).unwrap();
+}
+
+pub async fn collect_dates() {
+    let json_path = path!(*DICT_DIR / "history_date_samples.json");
+    let rp = RustyPipe::builder()
+        .storage_dir(path!(env!("CARGO_MANIFEST_DIR") / ".."))
+        .build()
+        .unwrap();
+
+    let mut res: CollectedDates = {
+        let json_file = File::open(&json_path).unwrap();
+        serde_json::from_reader(BufReader::new(json_file)).unwrap()
+    };
+
+    for lang in LANGUAGES {
+        println!("{lang}");
+        let history = rp.query().lang(lang).history().await.unwrap();
+        if history.items.len() < 3 {
+            panic!("{lang} empty history")
+        }
+
+        let entry = res.entry(lang).or_default();
+        entry.insert(
+            "tuesday".to_owned(),
+            history.items[0].playback_date_txt.clone().unwrap(),
+        );
+        entry.insert(
+            "0000-01-06".to_owned(),
+            history.items[1].playback_date_txt.clone().unwrap(),
+        );
+        entry.insert(
+            "2024-12-28".to_owned(),
+            history.items[15].playback_date_txt.clone().unwrap(),
+        );
+    }
+
+    let file = File::create(&json_path).unwrap();
     serde_json::to_writer_pretty(file, &res).unwrap();
 }
 
@@ -59,10 +100,10 @@ pub fn write_samples_to_dict() {
         let cd = &collected_dates[&lang];
         dict_entry
             .timeago_nd_tokens
-            .insert(util::filter_datestr(&cd.this_week), "0Wl".to_owned());
+            .insert(util::filter_datestr(&cd[THIS_WEEK]), "0Wl".to_owned());
         dict_entry
             .timeago_nd_tokens
-            .insert(util::filter_datestr(&cd.last_week), "1Wl".to_owned());
+            .insert(util::filter_datestr(&cd[LAST_WEEK]), "1Wl".to_owned());
     }
 
     util::write_dict(dict);

codegen/src/gen_dictionary.rs

@@ -45,7 +45,7 @@ pub fn generate_dictionary() {
 use crate::{
     model::AlbumType,
     param::Language,
-    util::timeago::{DateCmp, TaToken, TimeUnit},
+    util::timeago::{TaToken, TimeUnit},
 };
 
 /// Dictionary entry containing language-specific parsing information
@@ -57,14 +57,13 @@ pub(crate) struct Entry {
     /// Identifiers: `Y`(ear), `M`(month), `W`(eek), `D`(ay),
     /// `h`(our), `m`(inute), `s`(econd)
     pub timeago_tokens: phf::Map<&'static str, TaToken>,
-    /// Order in which to parse numeric date components. Formatted as
-    /// a string of date identifiers (Y, M, D).
+    /// True if the month has to be parsed before the day
     ///
     /// Examples:
     ///
-    /// - 03.01.2020 => `"DMY"`
-    /// - Jan 3, 2020 => `"DY"`
-    pub date_order: &'static [DateCmp],
+    /// - 03.01.2020 => DMY => false
+    /// - 01/03/2020 => MDY => true
+    pub month_before_day: bool,
     /// Tokens for parsing month names.
     ///
     /// Format: Parsed token -> Month number (starting from 1)
@@ -139,13 +138,6 @@ pub(crate) fn entry(lang: Language) -> Entry {
             };
         });
 
-        // Date order
-        let mut date_order = "&[".to_owned();
-        entry.date_order.chars().for_each(|c| {
-            write!(date_order, "DateCmp::{c}, ").unwrap();
-        });
-        date_order = date_order.trim_end_matches([' ', ',']).to_owned() + "]";
-
         // Number tokens
         let mut number_tokens = phf_codegen::Map::<&str>::new();
         entry.number_tokens.iter().for_each(|(txt, mag)| {
@@ -186,8 +178,8 @@ pub(crate) fn entry(lang: Language) -> Entry {
             .to_string()
             .replace('\n', "\n            ");
 
-        write!(code_timeago_tokens, "{} => Entry {{\n            timeago_tokens: {},\n            date_order: {},\n            months: {},\n            timeago_nd_tokens: {},\n            comma_decimal: {:?},\n            number_tokens: {},\n            number_nd_tokens: {},\n            album_types: {},\n            chan_prefix: {:?},\n            chan_suffix: {:?},\n        }},\n        ",
-        selector, code_ta_tokens, date_order, code_months, code_ta_nd_tokens, entry.comma_decimal, code_number_tokens, code_number_nd_tokens, code_album_types, entry.chan_prefix, entry.chan_suffix).unwrap();
+        write!(code_timeago_tokens, "{} => Entry {{\n            timeago_tokens: {},\n            month_before_day: {:?},\n            months: {},\n            timeago_nd_tokens: {},\n            comma_decimal: {:?},\n            number_tokens: {},\n            number_nd_tokens: {},\n            album_types: {},\n            chan_prefix: {:?},\n            chan_suffix: {:?},\n        }},\n        ",
+        selector, code_ta_tokens, entry.month_before_day, code_months, code_ta_nd_tokens, entry.comma_decimal, code_number_tokens, code_number_nd_tokens, code_album_types, entry.chan_prefix, entry.chan_suffix).unwrap();
     }
 
     code_timeago_tokens = code_timeago_tokens.trim_end().to_owned() + "\n    }\n}\n";

codegen/src/main.rs

@@ -32,6 +32,7 @@ enum Commands {
     CollectVideoDurations,
     CollectVideoDates,
     CollectHistoryDates,
+    CollectMusicHistoryDates,
     CollectChanPrefixes,
     ParsePlaylistDates,
     ParseHistoryDates,
@@ -74,6 +75,9 @@ async fn main() {
         Commands::CollectHistoryDates => {
             collect_history_dates::collect_dates().await;
         }
+        Commands::CollectMusicHistoryDates => {
+            collect_history_dates::collect_dates_music().await;
+        }
         Commands::CollectChanPrefixes => {
             collect_chan_prefixes::collect_chan_prefixes().await;
         }

codegen/src/model.rs

@@ -13,6 +13,13 @@ pub struct DictEntry {
     /// Should the language be parsed by character instead of by word?
     /// (e.g. Chinese/Japanese)
     pub by_char: bool,
+    /// True if the month has to be parsed before the day
+    ///
+    /// Examples:
+    ///
+    /// - 03.01.2020 => DMY => false
+    /// - 01/03/2020 => MDY => true
+    pub month_before_day: bool,
     /// Tokens for parsing timeago strings.
     ///
     /// Format: Parsed token -> \[Quantity\] Identifier