Update rust reporting implementation

reporting/rust/Cargo.lock

@@ -119,6 +119,12 @@ dependencies = [
  "syn",
 ]
 
+[[package]]
+name = "either"
+version = "1.8.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7fcaabb2fef8c910e7f4c7ce9f67a1283a1715879a7c230ca9d6d1ae31f16d91"
+
 [[package]]
 name = "iana-time-zone"
 version = "0.1.53"
@@ -143,6 +149,15 @@ dependencies = [
  "cxx-build",
 ]
 
+[[package]]
+name = "itertools"
+version = "0.10.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b0fd2260e829bddf4cb6ea802289de2f86d6a7a690192fbe91b3f46e0f2c8473"
+dependencies = [
+ "either",
+]
+
 [[package]]
 name = "js-sys"
 version = "0.3.61"
@@ -187,7 +202,9 @@ name = "nginx-parser"
 version = "0.1.0"
 dependencies = [
  "chrono",
+ "itertools",
  "regex",
+ "urlencoding",
 ]
 
 [[package]]
@@ -299,6 +316,12 @@ version = "0.1.10"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c0edd1e5b14653f783770bce4a4dabb4a5108a5370a5f5d8cfe8710c361f6c8b"
 
+[[package]]
+name = "urlencoding"
+version = "2.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e8db7427f936968176eaa7cdf81b7f98b980b18495ec28f1b5791ac3bfe3eea9"
+
 [[package]]
 name = "wasi"
 version = "0.10.0+wasi-snapshot-preview1"

reporting/rust/Cargo.toml

@@ -7,4 +7,6 @@ edition = "2021"
 
 [dependencies]
 chrono = "0.4.23"
+itertools = "0.10.5"
 regex = "1.7.1"
+urlencoding = "2.1.2"

reporting/rust/src/main.rs

-use chrono::DateTime;
+use chrono::{DateTime, FixedOffset};
+use itertools::Itertools;
 use regex::Regex;
 use std::collections::HashMap;
 use std::io::{self, BufRead};
 
+struct RecordInteraction {
+    views: i32,
+    exports: i32,
+    file_downloads: HashMap<String, i32>,
+}
+
+impl RecordInteraction {
+    fn zero() -> Self {
+        RecordInteraction {
+            views: 0,
+            exports: 0,
+            file_downloads: HashMap::new(),
+        }
+    }
+}
+
 fn main() {
     // TODO: argument parsing
+    let recid_prefix = "";
 
     let file_download_regex =
-        Regex::new(r"GET /records/(\w+\-\w+)/files/(.*?)(\?download=1|\s)").unwrap();
+        Regex::new(r"GET /records/(\w+\-\w+)/files/(.*?)(\?download=1|\?preview=\d|\s)").unwrap();
     let export_regex = Regex::new(r"GET /records/(\w+\-\w+)/export/").unwrap();
     let landing_page_regex = Regex::new(r"GET /records/(\w+\-\w+)\s").unwrap();
     let timestamp_regex =
         Regex::new(r"\[(\d{2}/\w{3}/\d{4}:\d{2}:\d{2}:\d{2}\s+.?\d{0,4})\]").unwrap();
 
-    let mut record_views = HashMap::new();
-    let mut file_downloads = HashMap::new();
-    let mut record_exports = HashMap::new();
+    let mut record_interactions = HashMap::new();
+    let mut start_timestamp: Option<DateTime<FixedOffset>> = None;
+    let mut end_timestamp: Option<DateTime<FixedOffset>> = None;
 
     for line in io::stdin().lock().lines() {
         let line = line.unwrap();
@@ -27,6 +45,10 @@ fn main() {
         if timestamp_regex.is_match(&line) {
             let capture = timestamp_regex.captures(&line).unwrap();
             timestamp = DateTime::parse_from_str(&capture[1], "%d/%b/%Y:%H:%M:%S %z").unwrap();
+            end_timestamp = Some(timestamp);
+            if start_timestamp == None {
+                start_timestamp = Some(timestamp);
+            }
         } else {
             continue;
         }
@@ -36,30 +58,53 @@ fn main() {
             let cap = landing_page_regex.captures(&line).unwrap();
             let recid = &cap[1];
 
-            let entry = record_views.entry(String::from(recid)).or_insert(0);
-            *entry += 1;
-
+            let entry = record_interactions
+                .entry(String::from(recid))
+                .or_insert(RecordInteraction::zero());
+            entry.views += 1;
         } else if file_download_regex.is_match(&line) {
             let cap = file_download_regex.captures(&line).unwrap();
             let recid = &cap[1];
-            let file = &cap[2];
-
-            let entry = file_downloads
-                .entry((String::from(recid), String::from(file)))
-                .or_insert(0);
-            *entry += 1;
+            let file = urlencoding::decode(&cap[2]).unwrap();
 
+            let entry = record_interactions
+                .entry(String::from(recid))
+                .or_insert(RecordInteraction::zero());
+            let count = entry.file_downloads.entry(String::from(file)).or_insert(0);
+            *count += 1;
         } else if export_regex.is_match(&line) {
             let cap = export_regex.captures(&line).unwrap();
             let recid = &cap[1];
 
-            let entry = record_exports.entry(String::from(recid)).or_insert(0);
-            *entry += 1;
+            let entry = record_interactions
+                .entry(String::from(recid))
+                .or_insert(RecordInteraction::zero());
+            entry.exports += 1;
         }
     }
 
-    // TODO: sort & pretty-print the results
-    println!("record views: {:?}", record_views);
-    println!("file downloads: {:?}", file_downloads);
-    println!("record exports: {:?}", record_exports);
+    let start = match start_timestamp {
+        Some(timestamp) => timestamp.to_rfc3339(),
+        None => panic!("No proper logs could be parsed!"),
+    };
+    let end = end_timestamp.unwrap().to_rfc3339();
+
+    println!("Start: {}", start);
+    println!("End:   {}", end);
+    println!("======={}", "=".repeat(end.len()));
+
+    // print the results
+    for recid in record_interactions.keys().sorted() {
+        let interactions = &record_interactions[recid];
+        println!(
+            "\n{}{} - {} views, {} exports",
+            recid_prefix, recid, interactions.views, interactions.exports
+        );
+        for file in interactions.file_downloads.keys().sorted() {
+            println!(
+                "\t{}: {} downloads",
+                file, interactions.file_downloads[file]
+            );
+        }
+    }
 }