File: si.py
   1 #!/usr/bin/python3
   2 
   3 # The MIT License (MIT)
   4 #
   5 # Copyright © 2025 pacman64
   6 #
   7 # Permission is hereby granted, free of charge, to any person obtaining a copy
   8 # of this software and associated documentation files (the “Software”), to deal
   9 # in the Software without restriction, including without limitation the rights
  10 # to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
  11 # copies of the Software, and to permit persons to whom the Software is
  12 # furnished to do so, subject to the following conditions:
  13 #
  14 # The above copyright notice and this permission notice shall be included in
  15 # all copies or substantial portions of the Software.
  16 #
  17 # THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  18 # IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  19 # FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
  20 # AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  21 # LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
  22 # OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  23 # SOFTWARE.
  24 
  25 
  26 from base64 import b64decode, b64encode
  27 from io import BytesIO
  28 from re import compile as compile_re, Pattern
  29 from socket import socket
  30 from sys import argv, exit, stderr, stdin
  31 from typing import Callable, Dict, List, Tuple
  32 from webbrowser import open_new_tab
  33 
  34 
  35 info = '''
  36 si [options...]
  37 
  38 
  39 Show It shows data read from standard-input, using your default web browser
  40 by auto-opening tabs, auto-detecing the data-format, and using a random port
  41 among those available.
  42 
  43 The localhost connection is available only until all data are transferred:
  44 this means refreshing your browser tab will lose your content, replacing it
  45 with a server-not-found message page.
  46 
  47 Dozens of common data-formats are recognized when piped from stdin, such as
  48 
  49   - HTML (web pages)
  50   - PDF
  51   - pictures (PNG, JPEG, SVG, WEBP, HEIC, AVIF, GIF, BMP)
  52   - audio (AAC, MP3, FLAC, WAV, AU, MIDI)
  53   - video (MP4, MOV, WEBM, MKV, AVI)
  54   - JSON
  55   - generic UTF-8 plain-text
  56 '''
  57 
  58 # handle standard help cmd-line options, quitting right away in that case
  59 if len(argv) > 1 and argv[1] in ('-h', '--h', '-help', '--help'):
  60     print(info.strip())
  61     exit(0)
  62 
  63 
  64 # hdr_dispatch groups format-description-groups by their first byte, thus
  65 # shortening total lookups for some data header: notice how the `ftyp` data
  66 # formats aren't handled here, since these can include any byte in parts of
  67 # their first few bytes
  68 hdr_dispatch: Dict[int, List[Tuple[bytes, str]]] = {
  69     0x00: [
  70         (b'\x00\x00\x01\xba', 'video/mpeg'),
  71         (b'\x00\x00\x01\xb3', 'video/mpeg'),
  72         (b'\x00\x00\x01\x00', 'image/x-icon'),
  73         (b'\x00\x00\x02\x00', 'image/vnd.microsoft.icon'), # .cur files
  74         (b'\x00asm', 'application/wasm'),
  75     ],
  76     0x1a: [(b'\x1a\x45\xdf\xa3', 'video/webm')], # matches general MKV format
  77     0x1f: [(b'\x1f\x8b\x08', 'application/gzip')],
  78     0x23: [
  79         (b'#! ', 'text/plain; charset=UTF-8'),
  80         (b'#!/', 'text/plain; charset=UTF-8'),
  81     ],
  82     0x25: [(b'%PDF', 'application/pdf'), (b'%!PS', 'application/postscript')],
  83     0x28: [(b'\x28\xb5\x2f\xfd', 'application/zstd')],
  84     0x2e: [(b'.snd', 'audio/basic')],
  85     0x47: [(b'GIF87a', 'image/gif'), (b'GIF89a', 'image/gif')],
  86     0x49: [
  87         # some MP3s start with an ID3 meta-data section
  88         (b'ID3\x02', 'audio/mpeg'),
  89         (b'ID3\x03', 'audio/mpeg'),
  90         (b'ID3\x04', 'audio/mpeg'),
  91         (b'II*\x00', 'image/tiff'),
  92     ],
  93     0x4d: [(b'MM\x00*', 'image/tiff'), (b'MThd', 'audio/midi')],
  94     0x4f: [(b'OggS', 'audio/ogg')],
  95     0x50: [(b'PK\x03\x04', 'application/zip')],
  96     0x53: [(b'SQLite format 3\x00', 'application/x-sqlite3')],
  97     0x63: [(b'caff\x00\x01\x00\x00', 'audio/x-caf')],
  98     0x66: [(b'fLaC', 'audio/x-flac')],
  99     0x7b: [(b'{\\rtf', 'application/rtf')],
 100     0x7f: [(b'\x7fELF', 'application/x-elf')],
 101     0x89: [(b'\x89PNG\x0d\x0a\x1a\x0a', 'image/png')],
 102     0xff: [
 103         (b'\xff\xd8\xff', 'image/jpeg'),
 104         # handle common ways MP3 data start
 105         (b'\xff\xf3\x48\xc4\x00', 'audio/mpeg'),
 106         (b'\xff\xfb', 'audio/mpeg'),
 107     ],
 108 }
 109 
 110 
 111 # ftyp_types helps func match_ftyp auto-detect MPEG-4-like formats
 112 ftyp_types: Tuple[Tuple[bytes, str]] = (
 113     (b'M4A ', 'audio/aac'),
 114     (b'M4A\x00', 'audio/aac'),
 115     (b'mp42', 'video/x-m4v'),
 116     (b'dash', 'audio/aac'),
 117     (b'isom', 'video/mp4'),
 118     # (b'isom', 'audio/aac'),
 119     (b'MSNV', 'video/mp4'),
 120     (b'qt  ', 'video/quicktime'),
 121     (b'heic', 'image/heic'),
 122     (b'avif', 'image/avif'),
 123 )
 124 
 125 # xmlish_heuristics helps func guess_mime auto-detect HTML, SVG, and XML
 126 xmlish_heuristics: Tuple[Tuple[bytes, str]] = (
 127     (b'<html>', 'text/html'), (b'<html ', 'text/html'),
 128     (b'<head>', 'text/html'), (b'<head ', 'text/html'),
 129     (b'<body>', 'text/html'), (b'<body ', 'text/html'),
 130     (b'<!DOCTYPE html>', 'text/html'), (b'<!DOCTYPE html ', 'text/html'),
 131     (b'<svg>', 'image/svg+xml'), (b'<svg ', 'image/svg+xml'),
 132     (b'<?xml>', 'application/xml'), (b'<?xml ', 'application/xml'),
 133 )
 134 
 135 # json_heuristics helps func guess_mime auto-detect JSON via regexes:
 136 # it's not perfect, but it seems effective-enough in practice
 137 json_heuristics: Tuple[Pattern] = (
 138     compile_re(b'''^\\s*\\{\\s*"'''),
 139     compile_re(b'''^\\s*\\{\\s*\\['''),
 140     compile_re(b'''^\\s*\\[\\s*"'''),
 141     compile_re(b'''^\\s*\\[\\s*\\{'''),
 142     compile_re(b'''^\\s*\\[\\s*\\['''),
 143 )
 144 
 145 
 146 def exact_match(header: bytes, maybe: bytes) -> bool:
 147     enough_bytes = len(header) >= len(maybe)
 148     return enough_bytes and all(x == y for x, y in zip(header, maybe))
 149 
 150 
 151 def match_riff(header: bytes) -> str:
 152     if len(header) < 12 or not header.startswith(b'RIFF'):
 153         return ''
 154 
 155     if header.find(b'WEBP', 8, 12) == 8:
 156         return 'image/webp'
 157     if header.find(b'WAVE', 8, 12) == 8:
 158         return 'audio/x-wav'
 159     if header.find(b'AVI ', 8, 12) == 8:
 160         return 'video/avi'
 161     return ''
 162 
 163 
 164 def match_form(header: bytes) -> str:
 165     if len(header) < 12 or not header.startswith(b'FORM'):
 166         return ''
 167 
 168     if header.find(b'AIFF', 8, 12) == 8:
 169         return 'audio/aiff'
 170     if header.find(b'AIFC', 8, 12) == 8:
 171         return 'audio/aiff'
 172     return ''
 173 
 174 
 175 def match_ftyp(header: bytes) -> str:
 176     # first 4 bytes can be anything, next 4 bytes must be ASCII 'ftyp'
 177     if len(header) < 12 or header.find(b'ftyp', 4, 8) != 4:
 178         return ''
 179 
 180     # next 4 bytes after the ASCII 'ftyp' declare the data-format
 181     for marker, mime in ftyp_types:
 182         if header.find(marker, 8, 12) == 8:
 183             return mime
 184 
 185     return ''
 186 
 187 
 188 def guess_mime(header: bytes, fallback: str) -> str:
 189     # no bytes, no match
 190     if len(header) == 0:
 191         return fallback
 192 
 193     # check the MPEG-4-like formats, the RIFF formats, and AIFF audio
 194     for f in (match_ftyp, match_riff, match_form):
 195         m = f(header)
 196         if m != '':
 197             return m
 198 
 199     # maybe it's a bitmap picture, which almost always has 40 on 15th byte
 200     if header.startswith(b'BM') and header.find(b'\x28', 8, 16) == 14:
 201         return 'image/x-bmp'
 202 
 203     # check general lookup-table
 204     if header[0] in hdr_dispatch:
 205         for maybe in hdr_dispatch[header[0]]:
 206             if exact_match(header, maybe[0]):
 207                 return maybe[1]
 208 
 209     # try HTML, SVG, and even generic XML
 210     if header.find(b'<', 0, 8) >= 0:
 211         for marker, mime in xmlish_heuristics:
 212             if header.find(marker, 0, 64) >= 0:
 213                 return mime
 214 
 215     # try some common cases for JSON
 216     for pattern in json_heuristics:
 217         if pattern.match(header):
 218             return 'application/json'
 219 
 220     # nothing matched
 221     return fallback
 222 
 223 
 224 def show_it(conn, start: bytes, rest) -> None:
 225     # handle base64-encoded data-URIs
 226     if start.startswith(b'data:'):
 227         i = start.find(b';base64,', 0, 64)
 228         if i > 0:
 229             mime_type = str(start[len('data:'):i], encoding='utf-8')
 230             encoded = BytesIO()
 231             encoded.write(start[i + len(';base64,'):])
 232             encoded.write(rest.read())
 233             decoded = b64decode(encoded.getvalue())
 234             encoded.close()
 235 
 236             inp = BytesIO(decoded)
 237             if mime_type == '':
 238                 start = inp.read(4096)
 239                 mime_type = guess_mime(start, 'text/plain; charset=UTF-8')
 240                 show_it_as(conn, start, inp, mime_type)
 241             else:
 242                 show_it_as(conn, bytes(), inp, mime_type)
 243             return
 244 
 245     mime_type = guess_mime(start, 'text/plain; charset=UTF-8')
 246     return show_it_as(conn, start, rest, mime_type)
 247 
 248 
 249 def show_it_as(conn, start: bytes, rest, mime_type: str) -> None:
 250     # read-ignore all client headers
 251     while True:
 252         if conn.recv(1024).endswith(b'\r\n\r\n'):
 253             break
 254 
 255     # web-browsers insist on auto-downloads when given wave or flac audio
 256     for e in ('audio/x-wav', 'audio/x-flac'):
 257         if e == mime_type:
 258             handle_sound_workaround(conn, mime_type, start, rest)
 259             return
 260 
 261     # web-browsers insist on auto-downloads when given bitmap pictures
 262     if mime_type == 'image/x-bmp':
 263         handle_image_workaround(conn, mime_type, start, rest)
 264         return
 265 
 266     conn.sendall(b'HTTP/1.1 200 OK\r\n')
 267     conn.sendall(bytes(f'Content-Type: {mime_type}\r\n', encoding='utf-8'))
 268     conn.sendall(b'Content-Disposition: inline\r\nConnection: close\r\n\r\n')
 269     conn.sendall(start)
 270     conn.sendfile(rest)
 271 
 272 
 273 def chunked_write(dest, src) -> None:
 274     while True:
 275         chunk = src.read(32 * 1024)
 276         if not chunk:
 277             return
 278         dest.write(chunk)
 279 
 280 
 281 def handle_sound_workaround(conn, mime: str, start: bytes, rest) -> None:
 282     def emit_inner_body() -> None:
 283         data = BytesIO()
 284         s = f'    <audio controls autofocus src="data:{mime};base64,'
 285         conn.sendall(bytes(s, encoding='utf-8'))
 286         data.write(start)
 287         chunked_write(data, rest)
 288         conn.sendall(b64encode(data.getvalue()))
 289         conn.sendall(b'"></audio>\n')
 290         data.close()
 291     handle_workaround(conn, 'Sound', emit_inner_body)
 292 
 293 
 294 def handle_image_workaround(conn, mime: str, start: bytes, rest) -> None:
 295     def emit_inner_body() -> None:
 296         data = BytesIO()
 297         s = bytes(f'    <img src="data:{mime};base64,', encoding='utf-8')
 298         conn.sendall(s)
 299         data.write(start)
 300         chunked_write(data, rest)
 301         conn.sendall(b64encode(data.getvalue()))
 302         conn.sendall(b'">\n')
 303         data.close()
 304     handle_workaround(conn, 'Bitmap Picture', emit_inner_body)
 305 
 306 
 307 start = '''
 308 <!DOCTYPE html>
 309 <html lang="en">
 310 <head>
 311     <meta charset="UTF-8">
 312     <link rel="icon" href="data:,">
 313     <meta name="viewport" content="width=device-width, initial-scale=1.0">
 314     <title>%%</title>
 315     <style>
 316         body { margin: auto; }
 317         audio { display: block; margin: auto; width: 90vw; }
 318         img { display: block; margin: auto; }
 319     </style>
 320 </head>
 321 <body>
 322 '''
 323 
 324 def handle_workaround(conn, title: str, handle_inner_body: Callable) -> None:
 325     s = bytes(start.replace('%%', title).lstrip('\n'), encoding='utf-8')
 326     conn.sendall(b'HTTP/1.1 200 OK\r\n')
 327     conn.sendall(b'Content-Type: text/html; charset=UTF-8\r\n')
 328     conn.sendall(b'Content-Disposition: inline\r\nConnection: close\r\n\r\n')
 329     conn.sendall(s)
 330     handle_inner_body()
 331     conn.sendall(b'</body>\n</html>\n')
 332 
 333 
 334 try:
 335     # opening socket on port 0 randomly picks an available port
 336     sock = socket()
 337     sock.bind(('localhost', 0))
 338     port = sock.getsockname()[1]
 339     sock.settimeout(10.0)
 340     # only handle one thing at a time, since it's a one-off server
 341     sock.listen(1)
 342 
 343     open_new_tab(f'http://localhost:{port}')
 344 
 345     # handle only a single request-response cycle
 346     conn, addr = sock.accept()
 347     show_it(conn, stdin.buffer.read(4096), stdin.buffer)
 348     conn.close()
 349 
 350     sock.close()
 351 except Exception as e:
 352     print(str(e), file=stderr)
 353     exit(1)