aboutsummaryrefslogtreecommitdiff
path: root/youtube_dl/jsinterp.py
diff options
context:
space:
mode:
authorPhilipp Hagemeister <phihag@phihag.de>2014-07-24 10:41:12 +0200
committerPhilipp Hagemeister <phihag@phihag.de>2014-07-24 10:41:14 +0200
commit825abb81759d76e53127644a45e1d6cb7ff4f654 (patch)
tree574e0f13d4be18b7fcc718ea01f32bb30f3d6877 /youtube_dl/jsinterp.py
parent8944ec0109b1e9c847f178755123d5453400dd50 (diff)
[jsinterp] Implement splice and general improvement
I still get 403s on YouTube though.
Diffstat (limited to 'youtube_dl/jsinterp.py')
-rw-r--r--youtube_dl/jsinterp.py86
1 files changed, 58 insertions, 28 deletions
diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py
index 13ad5ba1a..34e5307fa 100644
--- a/youtube_dl/jsinterp.py
+++ b/youtube_dl/jsinterp.py
@@ -1,5 +1,6 @@
from __future__ import unicode_literals
+import json
import re
from .utils import (
@@ -40,8 +41,9 @@ class JSInterpreter(object):
assign = lambda v: v
expr = stmt[len('return '):]
else:
- raise ExtractorError(
- 'Cannot determine left side of statement in %r' % stmt)
+ # Try interpreting it as an expression
+ expr = stmt
+ assign = lambda v: v
v = self.interpret_expression(expr, local_vars, allow_recursion)
return assign(v)
@@ -53,35 +55,62 @@ class JSInterpreter(object):
if expr.isalpha():
return local_vars[expr]
- m = re.match(r'^(?P<in>[a-z]+)\.(?P<member>.*)$', expr)
+ try:
+ return json.loads(expr)
+ except ValueError:
+ pass
+
+ m = re.match(
+ r'^(?P<var>[a-z]+)\.(?P<member>[^(]+)(?:\(+(?P<args>[^()]*)\))?$',
+ expr)
if m:
+ variable = m.group('var')
member = m.group('member')
- variable = m.group('in')
+ arg_str = m.group('args')
- if variable not in local_vars:
+ if variable in local_vars:
+ obj = local_vars[variable]
+ else:
if variable not in self._objects:
self._objects[variable] = self.extract_object(variable)
obj = self._objects[variable]
- key, args = member.split('(', 1)
- args = args.strip(')')
- argvals = [int(v) if v.isdigit() else local_vars[v]
- for v in args.split(',')]
- return obj[key](argvals)
-
- val = local_vars[variable]
- if member == 'split("")':
- return list(val)
- if member == 'join("")':
- return ''.join(val)
- if member == 'length':
- return len(val)
- if member == 'reverse()':
- return val[::-1]
- slice_m = re.match(r'slice\((?P<idx>.*)\)', member)
- if slice_m:
- idx = self.interpret_expression(
- slice_m.group('idx'), local_vars, allow_recursion - 1)
- return val[idx:]
+
+ if arg_str is None:
+ # Member access
+ if member == 'length':
+ return len(obj)
+ return obj[member]
+
+ assert expr.endswith(')')
+ # Function call
+ if arg_str == '':
+ argvals = tuple()
+ else:
+ argvals = tuple([
+ self.interpret_expression(v, local_vars, allow_recursion)
+ for v in arg_str.split(',')])
+
+ if member == 'split':
+ assert argvals == ('',)
+ return list(obj)
+ if member == 'join':
+ assert len(argvals) == 1
+ return argvals[0].join(obj)
+ if member == 'reverse':
+ assert len(argvals) == 0
+ return obj[::-1]
+ if member == 'slice':
+ assert len(argvals) == 1
+ return obj[argvals[0]:]
+ if member == 'splice':
+ assert isinstance(obj, list)
+ index, howMany = argvals
+ res = []
+ for i in range(index, min(index + howMany, len(obj))):
+ res.append(obj.pop(i))
+ return res
+
+ return obj[member](argvals)
m = re.match(
r'^(?P<in>[a-z]+)\[(?P<idx>.+)\]$', expr)
@@ -100,13 +129,14 @@ class JSInterpreter(object):
return a % b
m = re.match(
- r'^(?P<func>[a-zA-Z$]+)\((?P<args>[a-z0-9,]+)\)$', expr)
+ r'^(?P<func>[.a-zA-Z$]+)\((?P<args>[a-z0-9,]+)\)$', expr)
if m:
fname = m.group('func')
+ argvals = tuple([
+ int(v) if v.isdigit() else local_vars[v]
+ for v in m.group('args').split(',')])
if fname not in self._functions:
self._functions[fname] = self.extract_function(fname)
- argvals = [int(v) if v.isdigit() else local_vars[v]
- for v in m.group('args').split(',')]
return self._functions[fname](argvals)
raise ExtractorError('Unsupported JS expression %r' % expr)