From a5efd75f7954a64d2b8e1f3d4ecf6434774f6af3 Mon Sep 17 00:00:00 2001
From: Jordan <me@jordan.im>
Date: Thu, 24 Dec 2020 10:34:56 -0700
Subject: use curl for PDF downloads

---
 README   |  5 +++--
 pouch.sh | 26 +++++++++++++++++++-------
 2 files changed, 22 insertions(+), 9 deletions(-)

diff --git a/README b/README
index 39dac0f..059684b 100644
--- a/README
+++ b/README
@@ -1,7 +1,8 @@
 Usage: pouch.sh HTML_FILE_PATH
 
-      HTML_FILE is a pocket-exported document containing the set of URLs to save
+  HTML_FILE is a pocket-exported document containing the set of URLs to save
 
 Dependencies:
 
-      google-chrome 59+ (headless mode support)
+  google-chrome 59+ (headless mode support)
+  curl
diff --git a/pouch.sh b/pouch.sh
index 500a3a5..341e846 100755
--- a/pouch.sh
+++ b/pouch.sh
@@ -1,17 +1,19 @@
 #!/usr/bin/env bash
 
 PROGRAM="${0##*/}"
+AGENT="Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:84.0) Gecko/20100101 Firefox/84.0"
 
 cmd_usage() {
     cat >&2 <<-_EOF
-    Usage: $PROGRAM HTML_FILE_PATH
+Usage: $PROGRAM HTML_FILE_PATH
 
-      HTML_FILE is a pocket-exported document containing the set of URLs to save
+  HTML_FILE is a pocket-exported document containing the set of URLs to save
 
-    Dependencies:
+Dependencies:
 
-      google-chrome 59+ (headless mode support)
-    _EOF
+  google-chrome 59+ (headless mode support)
+  curl
+_EOF
 }
 
 get_urls() {
@@ -27,12 +29,22 @@ url_to_filename() {
 
 save() {
     read url
+
+    if [[ -f "$(url_to_filename $url).pdf" ]]; then
+        return 0
+    fi
+
+    content_type=$(curl -s -I -A "$AGENT" "$url" | grep --ignore-case '^content-type:' | cut -d' ' -f2)
     pdfname=$(url_to_filename $url).pdf
 
-    if [[ ! -f ./$pdfname ]]; then
+    if [[ $content_type == *"application/pdf"* ]]; then
         echo [+] $url...
-        google-chrome --headless --disable-gpu --print-to-pdf=$pdfname $url
+        curl -A "$AGENT" "$url" -o "$pdfname"
+    else
+        echo [+] $url...
+        google-chrome --headless --disable-gpu --print-to-pdf="$pdfname" "$url"
     fi
+
 }
 
 if [[ $# -eq 1 && ( $1 == --help || $1 == -h || $1 == help ) ]]; then
-- 
cgit v1.2.3-54-g00ecf