implement file name templating (title and timestamp)

Author: snshn

Cargo.toml

@@ -25,7 +25,7 @@ license = "CC0-1.0"
 [dependencies]
 atty = "0.2.14"  # Used for highlighting network errors
 base64 = "0.22.1"  # Used for integrity attributes
-chrono = "0.4.40"  # Used for formatting output timestamp
+chrono = "0.4.40"  # Used for formatting timestamps
 clap = { version = "4.5.32", features = ["derive"], optional = true }  # Used for processing CLI arguments
 cssparser = "0.34.0"  # Used for dealing with CSS
 encoding_rs = "0.8.35"  # Used for parsing and converting document charsets

src/core.rs

@@ -15,8 +15,8 @@ use url::Url;
 use crate::cache::Cache;
 use crate::cookies::Cookie;
 use crate::html::{
-    add_favicon, create_metadata_tag, get_base_url, get_charset, has_favicon, html_to_dom,
-    serialize_document, set_base_url, set_charset, walk_and_embed_assets,
+    add_favicon, create_metadata_tag, get_base_url, get_charset, get_title, has_favicon,
+    html_to_dom, serialize_document, set_base_url, set_charset, walk_and_embed_assets,
 };
 use crate::url::{clean_url, create_data_url, get_referer_url, parse_data_url, resolve_url};
 
@@ -122,7 +122,7 @@ pub fn create_monolithic_document(
     source: String,
     options: &Options,
     cache: &mut Option<Cache>,
-) -> Result<Vec<u8>, MonolithError> {
+) -> Result<(Vec<u8>, Option<String>), MonolithError> {
     // Check if source was provided
     if source.is_empty() {
         return Err(MonolithError::new("no target specified"));
@@ -235,7 +235,7 @@ pub fn create_monolithic_document(
                 if !media_type.eq_ignore_ascii_case("text/html")
                     && !media_type.eq_ignore_ascii_case("application/xhtml+xml")
                 {
-                    return Ok(retrieved_data);
+                    return Ok((retrieved_data, None));
                 }
 
                 if options
@@ -353,6 +353,8 @@ pub fn create_monolithic_document(
         dom = set_charset(dom, document_encoding.clone());
     }
 
+    let document_title: Option<String> = get_title(&dom.document);
+
     if options.output_format == MonolithOutputFormat::HTML {
         // Serialize DOM tree
         let mut result: Vec<u8> = serialize_document(dom, document_encoding, options);
@@ -364,9 +366,9 @@ pub fn create_monolithic_document(
             result.splice(0..0, metadata_comment.as_bytes().to_vec());
         }
 
-        Ok(result)
+        Ok((result, document_title))
     } else {
-        Ok(vec![])
+        Ok((vec![], document_title))
     }
 }
 

src/html.rs

@@ -124,7 +124,7 @@ pub fn compose_csp(options: &Options) -> String {
 }
 
 pub fn create_metadata_tag(url: &Url) -> String {
-    let timestamp = Utc::now().to_rfc3339_opts(SecondsFormat::Secs, true);
+    let datetime: &str = &Utc::now().to_rfc3339_opts(SecondsFormat::Secs, true);
     let mut clean_url: Url = clean_url(url.clone());
 
     // Prevent credentials from getting into metadata
@@ -141,7 +141,7 @@ pub fn create_metadata_tag(url: &Url) -> String {
         } else {
             "local source"
         },
-        timestamp,
+        datetime,
         env!("CARGO_PKG_NAME"),
         env!("CARGO_PKG_VERSION"),
     )
@@ -339,6 +339,18 @@ pub fn get_parent_node(child: &Handle) -> Handle {
     parent.and_then(|node| node.upgrade()).unwrap()
 }
 
+pub fn get_title(node: &Handle) -> Option<String> {
+    for title_node in find_nodes(node, vec!["html", "head", "title"]).iter() {
+        for child_node in title_node.children.borrow().iter() {
+            if let NodeData::Text { ref contents } = child_node.data {
+                return Some(contents.borrow().to_string());
+            }
+        }
+    }
+
+    None
+}
+
 pub fn has_favicon(handle: &Handle) -> bool {
     let mut found_favicon: bool = false;
 

src/main.rs

@@ -2,6 +2,7 @@ use std::fs;
 use std::io::{self, Error as IoError, Write};
 use std::process;
 
+use chrono::prelude::*;
 use clap::Parser;
 use tempfile::Builder;
 
@@ -118,11 +119,28 @@ enum Output {
 }
 
 impl Output {
-    fn new(destination: &str) -> Result<Output, IoError> {
+    fn new(destination: &str, document_title: &str) -> Result<Output, IoError> {
         if destination.is_empty() || destination.eq("-") {
             Ok(Output::Stdout(io::stdout()))
         } else {
-            Ok(Output::File(fs::File::create(destination)?))
+            let datetime: &str = &Utc::now().to_rfc3339_opts(SecondsFormat::Secs, true);
+            let final_destination = &destination
+                .replace("%timestamp%", &datetime.replace(':', "_"))
+                .replace(
+                    "%title%",
+                    document_title
+                        .to_string()
+                        .replace('/', "∕")
+                        .replace('\\', "⧵")
+                        .replace('<', "[")
+                        .replace('>', "]")
+                        .replace(':', "_")
+                        .replace('\"', "''")
+                        .replace('|', "_")
+                        .replace('?', "")
+                        .trim_start_matches('.'),
+                );
+            Ok(Output::File(fs::File::create(final_destination)?))
         }
     }
 
@@ -246,10 +264,13 @@ fn main() {
 
     // Retrieve target from source and output result
     match create_monolithic_document(cli.target, &options, &mut cache) {
-        Ok(result) => {
+        Ok((result, title)) => {
             // Define output
-            let mut output = Output::new(&destination.unwrap_or(String::new()))
-                .expect("could not prepare output");
+            let mut output = Output::new(
+                &destination.unwrap_or(String::new()),
+                &title.unwrap_or_default(),
+            )
+            .expect("could not prepare output");
 
             // Write result into STDOUT or file
             output.write(&result).expect("could not write output");