[dl:http] add MIME type and signature for .html files

2025-06-12 21:16:34 +02:00
parent e41c317039
commit 8b6bc54e95
2 changed files with 10 additions and 0 deletions
--- a/gallery_dl/downloader/http.py
+++ b/gallery_dl/downloader/http.py
@@ -469,12 +469,19 @@ MIME_TYPES = {
    "application/x-pdf": "pdf",
    "application/x-shockwave-flash": "swf",

+    "text/html": "html",
+
    "application/ogg": "ogg",
    # https://www.iana.org/assignments/media-types/model/obj
    "model/obj": "obj",
    "application/octet-stream": "bin",
 }

+
+def _signature_html(s):
+    return b"<!doctype html".startswith(s[:14].lower())
+
+
 # https://en.wikipedia.org/wiki/List_of_file_signatures
 SIGNATURE_CHECKS = {
    "jpg" : lambda s: s[0:3] == b"\xFF\xD8\xFF",
@@ -505,6 +512,8 @@ SIGNATURE_CHECKS = {
    "7z"  : lambda s: s[0:6] == b"\x37\x7A\xBC\xAF\x27\x1C",
    "pdf" : lambda s: s[0:5] == b"%PDF-",
    "swf" : lambda s: s[0:3] in (b"CWS", b"FWS"),
+    "html": _signature_html,
+    "htm" : _signature_html,
    "blend": lambda s: s[0:7] == b"BLENDER",
    # unfortunately the Wavefront .obj format doesn't have a signature,
    # so we check for the existence of Blender's comment
--- a/test/test_downloader.py
+++ b/test/test_downloader.py
@@ -369,6 +369,7 @@ SAMPLES = {
    ("heic", b"????ftypheis"),
    ("heic", b"????ftypheix"),
    ("svg" , b"<?xml"),
+    ("html", b"<!DOCTYPE html><html>...</html>"),
    ("ico" , b"\x00\x00\x01\x00"),
    ("cur" , b"\x00\x00\x02\x00"),
    ("psd" , b"8BPS"),