def extract_links_from_swf(file):
swf = swfparser.SWFParser(file)
urls = set()
for tag in swf.tags:
for text in analyze_tag(tag):
if looks_like_an_url(text):
urls.add(text)
return list(urls)
def scrap(fh, custom_order=None):
"""Get useful info from a program."""
swf = swfparser.SWFParser(fh)
# get the images
base = None
images = []
for tag in swf.tags:
if tag.name == 'JPEGTables':
base = tag.JPEGData
elif tag.name == 'DefineBits':
images.append((tag.CharacterID, tag.JPEGData))
elif tag.name == 'DefineBitsJPEG2':
images.append((tag.CharacterID, tag.ImageData))
images = [base + x[1] for x in sorted(images, reverse=True)]
# get the last DefineSprite
defsprite = None
for tag in swf.tags:
if tag.name == 'DefineSprite':
defsprite = tag
assert tag is not None, "DefineSprite not found"
# get the actions
doaction = defsprite.ControlTags[0]
for act in doaction.Actions:
if act.name == 'ActionConstantPool':
break
else:
if len(images) < 3:
# not enough images and no constant pool: a non-programs swf!
return []
raise ValueError("No ActionConstantPool found!")
# do some magic to retrieve the texts
cpe = _ConstantPoolExtractor(act.ConstantPool, doaction.Actions)
i = 0
all_vals = []
while True:
i += 1
name = 'titulo%d1' % i
occup = 'titulo%d2' % i
bio = 'htmlText'
date = 'titulo%d3' % i
vals = cpe.get(name, occup, bio, date)
if vals is None:
break
all_vals.append((vals[name], vals[occup], vals[bio], vals[date]))
items = []
for i, (name, occup, bio, date) in enumerate(all_vals):
date = _fix_date(date)
if date is None:
continue
occup = _fix_occup(occup)
bio = _fix_bio(bio)
name = _fix_name(name)
# use the corresponding image, or through the custom order
if custom_order is None:
idx = i
else:
try:
idx = custom_order.index(name)
except:
continue
image = images[idx]
ep = Episode(name=name, occup=occup, bio=bio, image=image, date=date)
items.append(ep)
return items