feat: add album/artist filtering to run command (#17)

* chore: remove stale TODO comment * feat: add album/artist filtering to run command * fix: remove unused DateTime import * chore: set MSRV to 1.82.0 * chore: update nix flake (for newer rustc)
Ovyerus · Dec 25, 2024 · ca9a5de · ca9a5de
1 parent 2a2baf7
commit ca9a5de
Show file tree

Hide file tree

Showing 6 changed files with 84 additions and 47 deletions.
diff --git a/Cargo.toml b/Cargo.toml
@@ -2,6 +2,7 @@
 name = "bandsnatch"
 version = "0.3.3"
 edition = "2021"
+rust-version = "1.82.0"
 description = "A CLI batch downloader for your Bandcamp collection"
 authors = ["Ashlynne Mitchell <ovy@ovyerus.com>"]
 license = "MIT"

diff --git a/flake.lock b/flake.lock
diff --git a/src/api/mod.rs b/src/api/mod.rs
@@ -97,8 +97,31 @@ impl Api {
         Ok(response)
     }
 
+    /// Filters the download map by optional artist or album filters.
+    fn filter_download_map<'a>(
+        unfiltered: Option<DownloadsMap>,
+        items: &'a Vec<&'a Item>,
+        album: Option<&String>,
+        artist: Option<&String>
+    ) -> DownloadsMap {
+        unfiltered
+            .iter()
+            .flatten()
+            .filter_map(|(id, url)| {
+                items.iter().find(|v| &format!("{}{}", v.sale_item_type, v.sale_item_id) == id)
+                    .filter(|item| {
+                        artist.is_none_or(|v| item.band_name.eq_ignore_ascii_case(v))
+                    })
+                    .filter(|item| {
+                        album.is_none_or(|v| item.item_title.eq_ignore_ascii_case(v))
+                    })
+                    .map(|_| (id.clone(), url.clone()))
+            })
+            .collect::<DownloadsMap>()
+    }
+
     /// Scrape a user's Bandcamp page to find download urls
-    pub fn get_download_urls(&self, name: &str) -> Result<BandcampPage, Box<dyn Error>> {
+    pub fn get_download_urls(&self, name: &str, artist: Option<&String>, album: Option<&String>) -> Result<BandcampPage, Box<dyn Error>> {
         debug!("`get_download_urls` for Bandcamp page '{name}'");
 
         let body = self.request(Method::GET, &Self::bc_path(name))?.text()?;
@@ -115,6 +138,8 @@ impl Api {
             .expect("Failed to deserialise collection page data blob.");
         debug!("Successfully fetched Bandcamp page, and found + deserialised data blob");
 
+        let items = fanpage_data.item_cache.collection.values().collect::<Vec<&Item>>();
+
         match fanpage_data.fan_data.is_own_page {
             Some(true) => (),
             _ => bail!(format!(
@@ -123,11 +148,7 @@ impl Api {
         }
 
         // TODO: make sure this exists
-        let mut collection = fanpage_data
-            .collection_data
-            .redownload_urls
-            .clone()
-            .unwrap();
+        let mut collection = Self::filter_download_map(fanpage_data.collection_data.redownload_urls.clone(), &items, album, artist);
 
         let skip_hidden_items = true;
         if skip_hidden_items {
@@ -142,7 +163,7 @@ impl Api {
                 // This should never be `None` thanks to the comparison above.
                 fanpage_data.collection_data.item_count.unwrap()
             );
-            let rest = self.get_rest_downloads_in_collection(&fanpage_data, "collection_items")?;
+            let rest = self.get_rest_downloads_in_collection(&fanpage_data, "collection_items", album, artist)?;
             collection.extend(rest);
         }
 
@@ -153,7 +174,7 @@ impl Api {
                 "Too many in `hidden_data`, and we're told not to skip, so we need to paginate ({} total)",
                 fanpage_data.hidden_data.item_count.unwrap()
             );
-            let rest = self.get_rest_downloads_in_collection(&fanpage_data, "hidden_items")?;
+            let rest = self.get_rest_downloads_in_collection(&fanpage_data, "hidden_items", album, artist)?;
             collection.extend(rest);
         }
 
@@ -171,6 +192,8 @@ impl Api {
         &self,
         data: &ParsedFanpageData,
         collection_name: &str,
+        album: Option<&String>,
+        artist: Option<&String>,
     ) -> Result<DownloadsMap, Box<dyn Error>> {
         debug!("Paginating results for {collection_name}");
         let collection_data = match collection_name {
@@ -199,8 +222,12 @@ impl Api {
                 .send()?
                 .json::<ParsedCollectionItems>()?;
 
-            trace!("Collected {} items", body.redownload_urls.clone().len());
-            collection.extend(body.redownload_urls);
+            let items = body.items.iter().by_ref().collect::<Vec<_>>();
+            let redownload_urls = Self::filter_download_map(Some(body.redownload_urls), &items, album, artist);
+            trace!("Collected {} items", redownload_urls.len());
+
+
+            collection.extend(redownload_urls);
             more_available = body.more_available;
             last_token = body.last_token;
         }

diff --git a/src/api/structs/digital_item.rs b/src/api/structs/digital_item.rs
@@ -1,6 +1,6 @@
 use crate::util::make_string_fs_safe;
 
-use chrono::{DateTime, Datelike, NaiveDateTime};
+use chrono::{Datelike, NaiveDateTime};
 use serde::{self, Deserialize};
 use std::{collections::HashMap, path::Path};
 

diff --git a/src/api/structs/mod.rs b/src/api/structs/mod.rs
@@ -17,7 +17,20 @@ pub struct ParsedFanpageData {
     pub collection_data: CollectionData,
     /// Data about items in the user's music collection that have been hidden.
     pub hidden_data: CollectionData,
-    // pub item_cache: ItemCache,
+    pub item_cache: ItemCache,
+}
+
+#[derive(Deserialize, Debug)]
+pub struct ItemCache {
+    pub collection: HashMap<String, Item>,
+}
+
+#[derive(Deserialize, Debug)]
+pub struct Item {
+    pub sale_item_id: u64,
+    pub sale_item_type: String,
+    pub band_name: String,
+    pub item_title: String,
 }
 
 #[derive(Deserialize, Debug)]
@@ -35,26 +48,13 @@ pub struct CollectionData {
     pub redownload_urls: Option<DownloadsMap>,
 }
 
-// #[derive(Deserialize, Debug)]
-// pub struct ItemCache {
-//     pub collection: HashMap<String, CachedItem>,
-//     pub hidden: HashMap<String, CachedItem>,
-// }
-
-// #[derive(Deserialize, Debug)]
-// pub struct CachedItem {
-//     #[serde(deserialize_with = "deserialize_string_from_number")]
-//     pub sale_item_id: String,
-//     pub band_name: String,
-//     pub item_title: String,
-// }
-
 /// Structure of the data returned from Bandcamp's collection API.
 #[derive(Deserialize, Debug)]
 pub struct ParsedCollectionItems {
     pub more_available: bool,
     pub last_token: String,
     pub redownload_urls: DownloadsMap,
+    pub items: Vec<Item>,
 }
 
 #[derive(Deserialize, Debug)]

diff --git a/src/cmds/run.rs b/src/cmds/run.rs
@@ -34,6 +34,12 @@ macro_rules! skip_err {
 
 #[derive(Debug, ClapArgs)]
 pub struct Args {
+    #[arg(long, env = "BS_ALBUM")]
+    album: Option<String>,
+
+    #[arg(long, env = "BS_ARTIST")]
+    artist: Option<String>,
+
     /// The audio format to download the files in.
     #[arg(short = 'f', long = "format", value_parser = PossibleValuesParser::new(FORMATS), env = "BS_FORMAT")]
     audio_format: String,
@@ -78,6 +84,8 @@ pub struct Args {
 
 pub fn command(
     Args {
+        album,
+        artist,
         audio_format,
         cookies,
         debug,
@@ -117,7 +125,7 @@ pub fn command(
         root.join("bandcamp-collection-downloader.cache"),
     )));
 
-    let download_urls = api.get_download_urls(&user)?.download_urls;
+    let download_urls = api.get_download_urls(&user, artist.as_ref(), album.as_ref())?.download_urls;
     let items = {
         // Lock gets freed after this block.
         let cache_content = cache.lock().unwrap().content()?;
@@ -139,8 +147,6 @@ pub fn command(
     let m = Arc::new(MultiProgress::new());
     let dry_run_results = Arc::new(Mutex::new(Vec::<String>::new()));
 
-    // TODO:  dry_run
-
     thread::scope(|scope| {
         for i in 0..jobs {
             let api = api.clone();