refactor(wip): use Metadata and Entry to form Inventory

also remove old functions

.config/cspell/project-dictionary.txt

@@ -12,6 +12,8 @@ diaryrs
 deserialised
 indoc
 mgmt
+tzdb
+serde
 
 # Diary Entries
 Shunde

Cargo.lock

@@ -200,6 +200,7 @@ dependencies = [
  "indoc",
  "regex",
  "serde",
+ "serde_yaml",
  "toml",
 ]
 
@@ -266,6 +267,12 @@ version = "1.70.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7943c866cc5cd64cbc25b2e01621d07fa8eb2a1a23160ee81ce38704e97b8ecf"
 
+[[package]]
+name = "itoa"
+version = "1.0.14"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d75a2a4b1b190afb6f5425f10f6a8f959d2ea0b9c2b1d79553551850539e4674"
+
 [[package]]
 name = "js-sys"
 version = "0.3.76"
@@ -418,6 +425,12 @@ version = "0.8.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "2b15c43186be67a4fd63bee50d0303afffcef381492ebe2c5d87f324e1b8815c"
 
+[[package]]
+name = "ryu"
+version = "1.0.18"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f3cb5ba0dc43242ce17de99c180e96db90b235b8a9fdc9543c96d2209116bd9f"
+
 [[package]]
 name = "serde"
 version = "1.0.216"
@@ -447,6 +460,19 @@ dependencies = [
  "serde",
 ]
 
+[[package]]
+name = "serde_yaml"
+version = "0.9.34+deprecated"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6a8b1a1a2ebf674015cc02edccce75287f1a0130d394307b36743c2f5d504b47"
+dependencies = [
+ "indexmap",
+ "itoa",
+ "ryu",
+ "serde",
+ "unsafe-libyaml",
+]
+
 [[package]]
 name = "shlex"
 version = "1.3.0"
@@ -516,6 +542,12 @@ version = "1.0.14"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "adb9e6ca4f869e1180728b7950e35922a7fc6397f7b641499e8f3ef06e50dc83"
 
+[[package]]
+name = "unsafe-libyaml"
+version = "0.2.11"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "673aac59facbab8a9007c7f6108d11f63b603f7cabff99fabf650fea5c32b861"
+
 [[package]]
 name = "utf8parse"
 version = "0.2.2"

Cargo.toml

@@ -11,4 +11,5 @@ iana-time-zone = "0.1.61"
 indoc = "2.0.5"
 regex = "1.11.1"
 serde = { version = "1.0.216", features = ["derive"] }
+serde_yaml = { version = "0.9.34" }
 toml = "0.8.19"

src/args.rs

@@ -35,7 +35,7 @@ pub struct Config {
 }
 
 impl Config {
-    const CONFIG_PATHS: [&str; 5] = [
+    const CONFIG_PATHS: [&'static str; 5] = [
         "diaryrs.toml",
         "diary.toml",
         "~/.config/diaryrs.toml",

src/main.rs

@@ -68,19 +68,25 @@ fn main() {
 
             let inventory = Inventory::take(&config).unwrap_or_fatal("Failed to take inventory");
 
+            let last_entry = inventory.entries.last();
+
             println!("*** Inventory ***");
             println!(
                 "For the {day_ordinal} of {month_name}, {year} in {}",
                 time.timezone
             );
             println!();
-            println!("You have {} diary entries so far.", inventory.entries);
+            match last_entry {
+                Some(entry) => println!("Your last entry was written in {} at {}", entry.timezone, entry.timestamp.to_string()),
+                None => println!("Your last entry is not known!")
+            }
+            println!("You have {} diary entries so far.", inventory.entries_count);
             println!(
                 "You have written {} words so far, with an average of {} words per entry.",
                 inventory.word_count, inventory.avg_word_count
             );
             println!();
-            println!("*** Happy journaling! ***")
+            println!("*** Happy journaling! ***");
         }
     }
 }

src/mgmt.rs

@@ -1,11 +1,14 @@
+use chrono::FixedOffset;
 use diaryrs::macros::UnwrapOrFatalAble;
 use regex::Regex;
+use serde::Deserialize;
 
 use crate::args::Config;
 
 use std::{fs, path};
 
 // Management for existing diary entries
+// TODO: Proper error handling for all the times we used `?`
 
 /// Overall stats and information about existing diary entries
 ///
@@ -13,7 +16,12 @@ use std::{fs, path};
 #[derive(Debug)]
 pub struct Inventory {
     /// Number of entries in the diary
-    pub entries: usize,
+    pub entries_count: usize,
+
+    /// A vector of [`Entry`] representing each diary entry
+    ///
+    /// Sorted oldest to newest based on enclosed timestamp
+    pub entries: Vec<Entry>,
 
     /// Total word count of all entries
     ///
@@ -24,6 +32,68 @@ pub struct Inventory {
     pub avg_word_count: f64,
 }
 
+/// Represents the front-matter metadata included within each entry. Does not necessarily have to be associated with any real diary entry.
+///
+/// "Lower-level" (more representative of what is on disk) than an [`Entry`]. For example, it does not do any validation or processing of the timezone
+#[derive(Debug, Deserialize)]
+pub struct Metadata {
+    timestamp: String,
+    timezone: String,
+}
+
+impl Metadata {
+    pub fn read_from_entry(contents: &str) -> Result<Self, Box<dyn std::error::Error>> {
+        // Match YAML section
+        let yaml_regex = Regex::new(r"(?U)---\n([\S\s]+)---\n")
+            .unwrap_or_fatal("Failed to compile YAML regex. Something is very wrong!");
+
+        let matches = yaml_regex
+            .captures(&contents)
+            .ok_or("YAML front-matter not found!")?;
+
+        let yaml_str = matches
+            .get(1)
+            .ok_or("YAML front-matter match not found!")?
+            .as_str();
+
+        let meta = serde_yaml::from_str::<Metadata>(yaml_str)?;
+
+        Ok(meta)
+    }
+}
+
+/// Represents a single diary entry (markdown file) and its associated stats/metadata
+///
+/// Does not include the actual contents of the entry, only stats
+///
+/// The offset of this entry can be retrieved from the timestamp. The tzdb-style timezone (e.g. `Australia/Sydney`)
+/// is separately specified. Using the offset is probably preferred as the offset should already account for regional daylight saving and other oddities.
+#[derive(Debug)]
+pub struct Entry {
+    pub path: path::PathBuf,
+
+    pub word_count: u64,
+    pub timestamp: chrono::DateTime<FixedOffset>,
+    pub timezone: chrono_tz::Tz,
+}
+
+impl Entry {
+    pub fn read(path: path::PathBuf) -> Result<Self, Box<dyn std::error::Error>> {
+        let contents = fs::read_to_string(&path)?;
+
+        let meta = Metadata::read_from_entry(&contents)?;
+        let timestamp = chrono::DateTime::parse_from_rfc3339(&meta.timestamp)?;
+        let timezone: chrono_tz::Tz = meta.timezone[0..meta.timezone.len() - 8].parse()?;
+
+        Ok(Self {
+            path,
+            word_count: count_words(&contents)?,
+            timestamp,
+            timezone,
+        })
+    }
+}
+
 impl Inventory {
     /// Take inventory of the diary entries
     ///
@@ -39,11 +109,21 @@ impl Inventory {
             .flatten()
             .collect();
 
-        let word_count = word_count_mds(&md_paths)?;
-        let entries_count = md_paths.len();
+        let mut entries = Vec::new();
+
+        for path in md_paths {
+            entries.push(Entry::read(path)?)
+        }
+
+        // Sort entry oldest to newest
+        entries.sort_by(|a, b| a.timestamp.cmp(&b.timestamp));
+
+        let word_count = entries.iter().fold(0, |acc, e| acc + e.word_count);
+        let entries_count = entries.len();
 
         Ok(Self {
-            entries: entries_count,
+            entries_count,
+            entries,
             word_count,
             avg_word_count: (word_count as f64 / entries_count as f64).round(),
         })
@@ -82,8 +162,7 @@ fn recurse_paths_md(dir: fs::DirEntry) -> Result<Vec<path::PathBuf>, Box<dyn std
     Ok(paths)
 }
 
-fn word_count_mds(paths: &Vec<path::PathBuf>) -> Result<u64, Box<dyn std::error::Error>> {
-    // temp regexes
+fn count_words(contents: &str) -> Result<u64, Box<dyn std::error::Error>> {
     let re_titles = Regex::new(r"(?m)^#{1,6} .+")
         .unwrap_or_fatal("Failed to compile title regex. Something is very wrong!");
     let re_comments = Regex::new(r"(?s-m)<!---?.+-->")
@@ -91,25 +170,17 @@ fn word_count_mds(paths: &Vec<path::PathBuf>) -> Result<u64, Box<dyn std::error:
     let re_images = Regex::new(r"!\[.*\]\(.+\)")
         .unwrap_or_fatal("Failed to compile image regex. Something is very wrong!");
 
-    let mut word_count = 0;
+    // Cut YAML header, comments and images
+    let contents = &re_titles.replace_all(
+        contents
+            .split("---")
+            .collect::<Vec<&str>>()
+            .pop()
+            .ok_or("No content found")?,
+        "",
+    );
+    let contents = &re_comments.replace_all(contents, "");
+    let contents = &re_images.replace_all(contents, "");
 
-    for path in paths {
-        let contents = fs::read_to_string(&path)?;
-
-        // Cut YAML header and comments
-        let contents = &re_titles.replace_all(
-            contents
-                .split("---")
-                .collect::<Vec<&str>>()
-                .pop()
-                .ok_or("No content found")?,
-            "",
-        );
-        let contents = &re_comments.replace_all(contents, "");
-        let contents = &re_images.replace_all(contents, "");
-
-        word_count += contents.split_whitespace().count();
-    }
-
-    Ok(word_count as u64)
+    Ok(contents.split_whitespace().count() as u64)
 }