| 1 |
# vim:fileencoding=utf-8 |
| 2 |
# License: GPL v3 Copyright: 2016, Kovid Goyal <kovid at kovidgoyal.net> |
| 3 |
from __python__ import hash_literals |
| 4 |
|
| 5 |
from elementmaker import E |
| 6 |
from encodings import base64decode, utf8_decode |
| 7 |
|
| 8 |
from dom import clear, remove_all_attributes |
| 9 |
from read_book.globals import runtime, ui_operations |
| 10 |
from read_book.settings import opts |
| 11 |
|
| 12 |
JSON_XHTML_MIMETYPE = 'application/calibre+xhtml+json' |
| 13 |
|
| 14 |
|
| 15 |
def decode_component(x): |
| 16 |
return utf8_decode(base64decode(x)) |
| 17 |
|
| 18 |
|
| 19 |
def decode_url(x): |
| 20 |
parts = x.split('#', 1) |
| 21 |
return decode_component(parts[0]), parts[1] or '' |
| 22 |
|
| 23 |
|
| 24 |
def create_link_pat(book): |
| 25 |
return RegExp(book.manifest.link_uid + r'\|([^|]+)\|', 'g') |
| 26 |
|
| 27 |
|
| 28 |
def load_resources(book, root_name, previous_resources, proceed): |
| 29 |
ans = Object.create(None) |
| 30 |
pending_resources = [root_name] |
| 31 |
link_pat = create_link_pat(book) |
| 32 |
|
| 33 |
def do_one(): |
| 34 |
name = pending_resources.shift() |
| 35 |
if not name: |
| 36 |
for k in previous_resources: |
| 37 |
v'delete previous_resources[k]' |
| 38 |
if book.manifest.files[root_name].has_maths: |
| 39 |
return load_mathjax(book, ans, proceed) |
| 40 |
return proceed(ans) |
| 41 |
if ans[name]: |
| 42 |
return setTimeout(do_one, 0) |
| 43 |
if previous_resources[name]: |
| 44 |
ans[name] = data = previous_resources[name] |
| 45 |
if jstype(data[0]) is 'string': |
| 46 |
find_virtualized_resources(data[0]) |
| 47 |
return setTimeout(do_one, 0) |
| 48 |
ui_operations.get_file(book, name, got_one) |
| 49 |
|
| 50 |
def got_one(data, name, mimetype): |
| 51 |
ans[name] = [data, mimetype] |
| 52 |
if jstype(data) is 'string' and book.manifest.files[name]?.is_virtualized: |
| 53 |
find_virtualized_resources(data) |
| 54 |
return setTimeout(do_one, 0) |
| 55 |
|
| 56 |
def find_virtualized_resources(text): |
| 57 |
seen = set() |
| 58 |
already_pending = {x.name for x in pending_resources} |
| 59 |
link_pat.lastIndex = 0 |
| 60 |
while True: |
| 61 |
m = link_pat.exec(text) |
| 62 |
if not m: |
| 63 |
break |
| 64 |
name = decode_url(m[1])[0] |
| 65 |
if name in seen or name in already_pending: |
| 66 |
continue |
| 67 |
seen.add(name) |
| 68 |
pending_resources.push(name) |
| 69 |
|
| 70 |
do_one() |
| 71 |
|
| 72 |
mathjax_data = None |
| 73 |
|
| 74 |
|
| 75 |
def load_mathjax(book, resource_data, proceed): |
| 76 |
if mathjax_data is None: |
| 77 |
ui_operations.get_mathjax_files(def (data): |
| 78 |
nonlocal mathjax_data |
| 79 |
mathjax_data = data |
| 80 |
resource_data['..mathjax-files..'] = data |
| 81 |
proceed(resource_data) |
| 82 |
) |
| 83 |
else: |
| 84 |
resource_data['..mathjax-files..'] = mathjax_data |
| 85 |
proceed(resource_data) |
| 86 |
|
| 87 |
|
| 88 |
def finalize_resources(book, root_name, resource_data): |
| 89 |
blob_url_map = Object.create(None) |
| 90 |
root_data = None |
| 91 |
link_pat = create_link_pat(book) |
| 92 |
mathjax = resource_data['..mathjax-files..'] |
| 93 |
v'delete resource_data["..mathjax-files.."]' |
| 94 |
|
| 95 |
# Resolve the non virtualized resources immediately |
| 96 |
for name in resource_data: |
| 97 |
data, mimetype = resource_data[name] |
| 98 |
if jstype(data) is not 'string': |
| 99 |
blob_url_map[name] = window.URL.createObjectURL(data) |
| 100 |
for name in blob_url_map: |
| 101 |
v'delete resource_data[name]' |
| 102 |
|
| 103 |
def add_virtualized_resource(name, text, mimetype): |
| 104 |
nonlocal root_data |
| 105 |
if name is root_name: |
| 106 |
root_data = JSON.parse(text) |
| 107 |
else: |
| 108 |
blob_url_map[name] = window.URL.createObjectURL(Blob([text], {'type': mimetype})) |
| 109 |
|
| 110 |
def replace_deps(text): |
| 111 |
replacements = [] |
| 112 |
unresolved_deps = set() |
| 113 |
link_pat.lastIndex = 0 |
| 114 |
while True: |
| 115 |
m = link_pat.exec(text) |
| 116 |
if not m: |
| 117 |
break |
| 118 |
dname, frag = decode_url(m[1]) |
| 119 |
if blob_url_map[dname]: |
| 120 |
rtext = blob_url_map[dname] |
| 121 |
if frag: |
| 122 |
rtext += '#' + frag |
| 123 |
replacements.push([m.index, m[0].length, rtext]) |
| 124 |
else: |
| 125 |
unresolved_deps.add(dname) |
| 126 |
for index, sz, repl in reversed(replacements): |
| 127 |
text = text[:index] + repl + text[index + sz:] |
| 128 |
return unresolved_deps, text |
| 129 |
|
| 130 |
unresolved_deps_map = {} |
| 131 |
|
| 132 |
def has_unresolvable_deps(name): |
| 133 |
deps = unresolved_deps_map[name] |
| 134 |
if not deps or not deps.length: |
| 135 |
return False |
| 136 |
for x in deps: |
| 137 |
if not blob_url_map[x]: |
| 138 |
return True |
| 139 |
return False |
| 140 |
|
| 141 |
while True: |
| 142 |
resolved = [] |
| 143 |
num = 0 |
| 144 |
for name in resource_data: |
| 145 |
if not blob_url_map[name]: |
| 146 |
num += 1 |
| 147 |
text, mimetype = resource_data[name] |
| 148 |
if not has_unresolvable_deps(name): |
| 149 |
unresolved_deps, text = replace_deps(text) |
| 150 |
unresolved_deps_map[name] = unresolved_deps |
| 151 |
if not unresolved_deps.length: |
| 152 |
add_virtualized_resource(name, text, mimetype) |
| 153 |
resolved.push(name) |
| 154 |
if not num: |
| 155 |
break |
| 156 |
if not resolved.length: |
| 157 |
unresolved = [name for name in resource_data if not blob_url_map[name]] |
| 158 |
print('ERROR: Could not resolve all dependencies of {} because of a cyclic dependency. Remaining deps: {}'.format(root_name, unresolved)) |
| 159 |
# Add the items anyway, without resolving remaining deps |
| 160 |
for name in resource_data: |
| 161 |
if not blob_url_map[name]: |
| 162 |
text, mimetype = resource_data[name] |
| 163 |
text = replace_deps(text)[1] |
| 164 |
add_virtualized_resource(name, text, mimetype) |
| 165 |
break |
| 166 |
for name in resolved: |
| 167 |
v'delete resource_data[name]' |
| 168 |
|
| 169 |
return root_data, mathjax, blob_url_map |
| 170 |
|
| 171 |
js_types = {k: True for k in 'text/javascript text/ecmascript application/javascript application/ecmascript'.split(' ')} |
| 172 |
resource_tag_names = { |
| 173 |
'script': 'src', |
| 174 |
'link': 'href', |
| 175 |
'img': 'src', |
| 176 |
'image': 'xlink:href', |
| 177 |
} |
| 178 |
ns_rmap = { |
| 179 |
'http://www.w3.org/2000/svg': 'svg', |
| 180 |
'http://www.w3.org/1999/xlink': 'xlink', |
| 181 |
'http://www.w3.org/1998/Math/MathML': 'math', |
| 182 |
'http://www.w3.org/XML/1998/namespace': 'xml', |
| 183 |
'http://www.idpf.org/2007/ops': 'epub', |
| 184 |
} |
| 185 |
ns_count = 0 |
| 186 |
hide_tooltips = False |
| 187 |
|
| 188 |
|
| 189 |
def get_prefix(ns): |
| 190 |
nonlocal ns_count |
| 191 |
ans = ns_rmap[ns] |
| 192 |
if not ans: |
| 193 |
ns_rmap[ns] = ans = 'ns' + ns_count |
| 194 |
ns_count += 1 |
| 195 |
return ans + ':' |
| 196 |
|
| 197 |
|
| 198 |
def apply_attributes(src, elem, ns_map): |
| 199 |
attributes = src.a |
| 200 |
if not attributes: |
| 201 |
return |
| 202 |
for a in attributes: |
| 203 |
if a[2]: |
| 204 |
ns = ns_map[a[2]] |
| 205 |
elem.setAttributeNS(ns, get_prefix(ns) + a[0], a[1]) |
| 206 |
else: |
| 207 |
name = a[0] |
| 208 |
if hide_tooltips and (name is 'title' or name is 'alt'): |
| 209 |
continue |
| 210 |
elem.setAttribute(name, a[1]) |
| 211 |
|
| 212 |
|
| 213 |
def is_loadable_link(attributes): |
| 214 |
for a in attributes: |
| 215 |
if a[0].toLowerCase() is 'rel' and a[1]: |
| 216 |
for x in a[1].split(' '): |
| 217 |
if x.toLowerCase() is 'stylesheet': |
| 218 |
return True |
| 219 |
return False |
| 220 |
|
| 221 |
|
| 222 |
def process_stack(stack, tag_map, ns_map, load_required, onload): |
| 223 |
while stack.length: |
| 224 |
node, parent = stack.pop() |
| 225 |
if tag_map: |
| 226 |
tag_id = node[0] |
| 227 |
src = tag_map[tag_id] |
| 228 |
else: |
| 229 |
src = node |
| 230 |
tag_id = v'process_stack.tag_id++' |
| 231 |
if src.s: |
| 232 |
if src.n: |
| 233 |
elem = document.createElementNS(ns_map[src.s], src.n) |
| 234 |
else: |
| 235 |
if src.l: |
| 236 |
parent.appendChild(document.createTextNode(src.l)) |
| 237 |
continue |
| 238 |
else: |
| 239 |
elem = document.createElement(src.n) |
| 240 |
loadable = False |
| 241 |
attr = resource_tag_names[src.n] |
| 242 |
if attr: |
| 243 |
if attr.indexOf(':') != -1: |
| 244 |
attr = attr.replace('xlink:', '') |
| 245 |
if src.a: |
| 246 |
for a in src.a: |
| 247 |
if a[0] is attr: |
| 248 |
loadable = is_loadable_link(src.a) if src.n is 'link' else True |
| 249 |
break |
| 250 |
if loadable: |
| 251 |
load_required.add(tag_id) |
| 252 |
load_callback = onload.bind(tag_id) |
| 253 |
elem.addEventListener('load', load_callback) |
| 254 |
elem.addEventListener('error', load_callback) |
| 255 |
|
| 256 |
apply_attributes(src, elem, ns_map) |
| 257 |
parent.appendChild(elem) |
| 258 |
if src.x: |
| 259 |
if src.n is 'script' and js_types[(elem.getAttribute('type') or 'text/javascript').toLowerCase()] is True: |
| 260 |
elem.text = src.x |
| 261 |
else: |
| 262 |
elem.appendChild(document.createTextNode(src.x)) |
| 263 |
if src.l: |
| 264 |
parent.appendChild(document.createTextNode(src.l)) |
| 265 |
if tag_map: |
| 266 |
for v'var i = node.length - 1; i >= 1; i--': # noqa: unused-local |
| 267 |
stack.push([node[i], elem]) |
| 268 |
elif node.c: |
| 269 |
for v'var i = node.c.length; i-- > 0;': # noqa: unused-local |
| 270 |
stack.push([node.c[i], elem]) |
| 271 |
|
| 272 |
|
| 273 |
def unserialize_html(serialized_data, proceed, postprocess_dom, root_name): |
| 274 |
nonlocal hide_tooltips |
| 275 |
hide_tooltips = opts.hide_tooltips |
| 276 |
if serialized_data.tag_map: |
| 277 |
return unserialize_html_legacy( |
| 278 |
serialized_data, |
| 279 |
proceed, |
| 280 |
postprocess_dom, |
| 281 |
root_name |
| 282 |
) |
| 283 |
html = serialized_data.tree |
| 284 |
ns_map = serialized_data.ns_map |
| 285 |
remove_all_attributes( |
| 286 |
document.documentElement, |
| 287 |
document.head, |
| 288 |
document.body |
| 289 |
) |
| 290 |
clear(document.head, document.body) |
| 291 |
apply_attributes(html, document.documentElement, ns_map) |
| 292 |
# hide browser scrollbars while loading since they will anyway be hidden |
| 293 |
# after loading and this prevents extra layouts |
| 294 |
document.head.appendChild(E.style(type='text/css', 'html::-webkit-scrollbar, body::-webkit-scrollbar { display: none !important }')) |
| 295 |
if runtime.is_standalone_viewer and root_name: |
| 296 |
if root_name.indexOf('/') > -1: |
| 297 |
base = window.location.pathname.rpartition('/')[0] |
| 298 |
base = f'{window.location.protocol}//{window.location.hostname}{base}/' + root_name |
| 299 |
document.head.appendChild(E.base(href=base)) |
| 300 |
|
| 301 |
# Default stylesheet |
| 302 |
if not runtime.is_standalone_viewer: |
| 303 |
# for the standalone viewer the default font family is set |
| 304 |
# in the viewer settings |
| 305 |
document.head.appendChild(E.style(type='text/css', 'html {{ font-family: {} }}'.format(window.default_font_family or 'sans-serif'))) |
| 306 |
|
| 307 |
load_required = set() |
| 308 |
proceeded = False |
| 309 |
hang_timeout = 5 |
| 310 |
|
| 311 |
def hangcheck(): |
| 312 |
nonlocal proceeded |
| 313 |
if not proceeded: |
| 314 |
proceeded = True |
| 315 |
print(f'WARNING: All resources did not load in {hang_timeout} seconds, proceeding anyway ({load_required.length} resources left)') |
| 316 |
proceed() |
| 317 |
|
| 318 |
def onload(): |
| 319 |
nonlocal proceeded |
| 320 |
load_required.discard(this) |
| 321 |
if not load_required.length and not proceeded: |
| 322 |
proceeded = True |
| 323 |
proceed() |
| 324 |
|
| 325 |
def process_children(node, parent): |
| 326 |
if not node.c: |
| 327 |
return |
| 328 |
stack = [] |
| 329 |
for v'var i = node.c.length; i-- > 0;': # noqa: unused-local |
| 330 |
child = v'node.c[i]' |
| 331 |
if child.n is not 'meta' and child.n is not 'base': |
| 332 |
stack.push([child, parent]) |
| 333 |
process_stack(stack, None, ns_map, load_required, onload) |
| 334 |
|
| 335 |
body_done = False |
| 336 |
process_stack.tag_id = 1 |
| 337 |
for child in html.c: |
| 338 |
if child.n is 'head': |
| 339 |
process_children(child, document.head) |
| 340 |
elif child.n is 'body': |
| 341 |
if not document.body: |
| 342 |
document.documentElement.appendChild(document.createElement('body')) |
| 343 |
if not body_done: |
| 344 |
body_done = True |
| 345 |
apply_attributes(child, document.body, ns_map) |
| 346 |
if child.x: |
| 347 |
document.body.appendChild(document.createTextNode(child.x)) |
| 348 |
process_children(child, document.body) |
| 349 |
|
| 350 |
if postprocess_dom: |
| 351 |
postprocess_dom() |
| 352 |
ev = document.createEvent('Event') |
| 353 |
ev.initEvent('DOMContentLoaded', True, True) |
| 354 |
document.dispatchEvent(ev) |
| 355 |
if load_required.length: |
| 356 |
setTimeout(hangcheck, hang_timeout * 1000) |
| 357 |
else: |
| 358 |
proceeded = True |
| 359 |
proceed() |
| 360 |
|
| 361 |
|
| 362 |
def unserialize_html_legacy(serialized_data, proceed, postprocess_dom, root_name): |
| 363 |
tag_map = serialized_data.tag_map |
| 364 |
tree = serialized_data.tree |
| 365 |
ns_map = serialized_data.ns_map |
| 366 |
html = tag_map[0] |
| 367 |
remove_all_attributes(document.documentElement) |
| 368 |
apply_attributes(html, document.documentElement, ns_map) |
| 369 |
head, body = tree[1], tree[2] |
| 370 |
clear(document.head, document.body) |
| 371 |
remove_all_attributes(document.head, document.body) |
| 372 |
# hide browser scrollbars while loading since they will anyway be hidden |
| 373 |
# after loading and this prevents extra layouts |
| 374 |
document.head.appendChild(E.style(type='text/css', 'html::-webkit-scrollbar, body::-webkit-scrollbar { display: none !important }')) |
| 375 |
if runtime.is_standalone_viewer and root_name: |
| 376 |
if root_name.indexOf('/') > -1: |
| 377 |
base = window.location.pathname.rpartition('/')[0] |
| 378 |
base = f'{window.location.protocol}//{window.location.hostname}{base}/' + root_name |
| 379 |
document.head.appendChild(E.base(href=base)) |
| 380 |
|
| 381 |
# Default stylesheet |
| 382 |
if not runtime.is_standalone_viewer: |
| 383 |
# for the standalone viewer the default font family is set |
| 384 |
# in the viewer settings |
| 385 |
document.head.appendChild(E.style(type='text/css', 'html {{ font-family: {} }}'.format(window.default_font_family or 'sans-serif'))) |
| 386 |
load_required = set() |
| 387 |
proceeded = False |
| 388 |
hang_timeout = 5 |
| 389 |
|
| 390 |
def hangcheck(): |
| 391 |
nonlocal proceeded |
| 392 |
if not proceeded: |
| 393 |
proceeded = True |
| 394 |
print(f'WARNING: All resources did not load in {hang_timeout} seconds, proceeding anyway ({load_required.length} resources left)') |
| 395 |
proceed() |
| 396 |
|
| 397 |
def onload(): |
| 398 |
nonlocal proceeded |
| 399 |
load_required.discard(this) |
| 400 |
if not load_required.length and not proceeded: |
| 401 |
proceeded = True |
| 402 |
proceed() |
| 403 |
|
| 404 |
stack = [] |
| 405 |
for v'var i = head.length - 1; i >= 1; i--': |
| 406 |
stack.push([head[i], document.head]) |
| 407 |
process_stack(stack, tag_map, ns_map, load_required, onload) |
| 408 |
bnode = tag_map[body[0]] |
| 409 |
apply_attributes(bnode, document.body, ns_map) |
| 410 |
if bnode.x: |
| 411 |
document.body.appendChild(document.createTextNode(bnode.x)) |
| 412 |
for v'var i = body.length - 1; i >= 1; i--': |
| 413 |
stack.push([body[i], document.body]) |
| 414 |
process_stack(stack, tag_map, ns_map, load_required, onload) |
| 415 |
if postprocess_dom: |
| 416 |
postprocess_dom() |
| 417 |
ev = document.createEvent('Event') |
| 418 |
ev.initEvent('DOMContentLoaded', True, True) |
| 419 |
document.dispatchEvent(ev) |
| 420 |
if load_required.length: |
| 421 |
setTimeout(hangcheck, hang_timeout * 1000) |
| 422 |
else: |
| 423 |
proceeded = True |
| 424 |
proceed() |
| 425 |
|
| 426 |
|
| 427 |
def text_from_serialized_html(data, get_anchor_offset_map): |
| 428 |
serialized_data = JSON.parse(data) |
| 429 |
tag_map = serialized_data.tag_map |
| 430 |
ans = [] |
| 431 |
no_visit = {'script': True, 'style': True, 'title': True, 'head': True} |
| 432 |
ignore_text = { |
| 433 |
'img': True, |
| 434 |
'math': True, |
| 435 |
'rt': True, |
| 436 |
'rp': True, |
| 437 |
'rtc': True, |
| 438 |
} |
| 439 |
ignore_text |
| 440 |
if tag_map: |
| 441 |
stack = [[serialized_data.tree[2], False]] |
| 442 |
else: |
| 443 |
stack = [] |
| 444 |
for child in serialized_data.tree.c: |
| 445 |
if child.n is 'body': |
| 446 |
stack.push([child, False]) |
| 447 |
anchor_offset_map = {} |
| 448 |
text_pos = 0 |
| 449 |
while stack.length: |
| 450 |
node, text_ignored_in_parent = stack.pop() |
| 451 |
if jstype(node) is 'string': |
| 452 |
ans.push(node) |
| 453 |
text_pos += node.length |
| 454 |
continue |
| 455 |
src = tag_map[node[0]] if tag_map else node |
| 456 |
if get_anchor_offset_map and src.a: |
| 457 |
for v'var i = 0; i < src.a.length; i++': |
| 458 |
x = src.a[i] |
| 459 |
if x[0] is 'id': |
| 460 |
aid = x[1] |
| 461 |
if jstype(anchor_offset_map[aid]) is not 'number': |
| 462 |
anchor_offset_map[aid] = text_pos |
| 463 |
if no_visit[src.n]: |
| 464 |
continue |
| 465 |
ignore_text_in_node_and_children = text_ignored_in_parent or v'!!ignore_text[src.n]' |
| 466 |
if not ignore_text_in_node_and_children and src.x: |
| 467 |
ans.push(src.x) |
| 468 |
text_pos += src.x.length |
| 469 |
if not text_ignored_in_parent and src.l: |
| 470 |
stack.push([src.l, ignore_text_in_node_and_children]) |
| 471 |
if tag_map: |
| 472 |
for v'var i = node.length - 1; i >= 1; i--': |
| 473 |
stack.push([node[i], ignore_text_in_node_and_children]) |
| 474 |
else: |
| 475 |
if src.c: |
| 476 |
for v'var i = src.c.length; i-- > 0;': |
| 477 |
stack.push([src.c[i], ignore_text_in_node_and_children]) |
| 478 |
ans = ans.join('') |
| 479 |
if get_anchor_offset_map: |
| 480 |
return ans, anchor_offset_map |
| 481 |
return ans |