rewrite refactoring:

- rewrite headers after content to ensure content-length/content-encoding rewritten if content modified
- header rewriter: remove proxyrewriter, set default rule to 'prefix' or 'keep' if url rewriting or not
- set is_content_rw if record.content_stream(), assume content is modified
- add BufferedRewriter as base for dash, hls, amf rewriting which processes the full stream
- should_rw_content() determines if should attempt content rewriting
- support banner-only insert mode: added HTMLInsertOnlyRewriter, enable if no custom JS rules
- test: enable banner-only test mode

pywb/rewrite/content_rewriter.py

@@ -7,6 +7,7 @@ from warcio.utils import to_native_str
 
 import re
 import webencodings
+import tempfile
 
 from pywb.webagg.utils import StreamIter, BUFF_SIZE
 from pywb.rewrite.cookie_rewriter import ExactPathCookieRewriter
@@ -78,11 +79,16 @@ class BaseContentRewriter(object):
     def create_rewriter(self, text_type, rule, rwinfo, cdx, head_insert_func=None):
         rw_type, rw_class = self.get_rw_class(rule, text_type, rwinfo)
 
-        if rw_type in ('js', 'js_proxy'):
+        if rw_type in ('js', 'js-proxy'):
             extra_rules = []
             if 'js_regex_func' in rule:
                 extra_rules = rule['js_regex_func'](rwinfo.url_rewriter)
 
+            # if js-proxy and no rules, default to none
+            # js rewriting in proxy only if extra rules apply
+            if rw_type == 'js-proxy' and not extra_rules:
+                return None
+
             return rw_class(rwinfo.url_rewriter, extra_rules)
 
         elif rw_type != 'html':
@@ -94,6 +100,10 @@ class BaseContentRewriter(object):
         js_rewriter = self.create_rewriter('js', rule, rwinfo, cdx)
         css_rewriter = self.create_rewriter('css', rule, rwinfo, cdx)
 
+        # if no js rewriter, then do banner insert only
+        if not js_rewriter:
+            rw_class = self.all_rewriters.get('html-banner-only')
+
         rw = rw_class(rwinfo.url_rewriter,
                       js_rewriter=js_rewriter,
                       css_rewriter=css_rewriter,
@@ -140,33 +150,28 @@ class BaseContentRewriter(object):
         return charset
 
     def rewrite_headers(self, rwinfo):
-        if rwinfo.is_url_rw():
-            header_rw_name = 'header'
-        else:
-            header_rw_name = 'header-proxy'
-
-        header_rw_class = self.all_rewriters.get(header_rw_name)
-        rwinfo.rw_http_headers = header_rw_class(rwinfo)()
+        header_rw_class = self.all_rewriters.get('header')
+        return header_rw_class(rwinfo)()
 
     def __call__(self, record, url_rewriter, cookie_rewriter,
                  head_insert_func=None,
                  cdx=None):
 
         rwinfo = RewriteInfo(record, self.get_rewrite_types(), url_rewriter, cookie_rewriter)
-
-        self.rewrite_headers(rwinfo)
-
         content_rewriter = None
-        if rwinfo.is_content_rw():
+
+        if rwinfo.should_rw_content():
             rule = self.get_rule(cdx)
             content_rewriter = self.create_rewriter(rwinfo.text_type, rule, rwinfo, cdx, head_insert_func)
 
         if content_rewriter:
             gen = content_rewriter(rwinfo)
         else:
-            gen = StreamIter(rwinfo.content_stream)
+            gen = StreamIter(rwinfo.record.raw_stream)
 
-        return rwinfo.rw_http_headers, gen, (content_rewriter != None)
+        rw_http_headers = self.rewrite_headers(rwinfo)
+
+        return rw_http_headers, gen, (content_rewriter != None)
 
     def init_js_regexs(self, regexs):
         raise NotImplemented()
@@ -175,10 +180,34 @@ class BaseContentRewriter(object):
         raise NotImplemented()
 
 
+# ============================================================================
+class BufferedRewriter(object):
+    def __init__(self, url_rewriter=None):
+        self.url_rewriter = url_rewriter
+
+    def __call__(self, rwinfo):
+        stream_buffer = tempfile.SpooledTemporaryFile(BUFF_SIZE * 4)
+
+        with closing(rwinfo.content_stream) as fh:
+            while True:
+                buff = fh.read()
+                if not buff:
+                    break
+
+                stream_buffer.write(buff)
+
+        stream_buffer.seek(0)
+        return StreamIter(self.rewrite_stream(stream_buffer))
+
+    def rewrite_stream(self, stream):
+        raise NotImplemented('implement in subclass')
+
+
 # ============================================================================
 class StreamingRewriter(object):
-    def __init__(self):
-        self.align_to_line = True
+    def __init__(self, url_rewriter, align_to_line=True):
+        self.url_rewriter = url_rewriter
+        self.align_to_line = align_to_line
 
     def __call__(self, rwinfo):
         gen = self.rewrite_text_stream_to_gen(rwinfo.content_stream,
@@ -233,8 +262,8 @@ class RewriteInfo(object):
     def __init__(self, record, rewrite_types, url_rewriter, cookie_rewriter):
         self.record = record
 
-        self.rw_http_headers = record.http_headers
-        self.content_stream = record.content_stream()
+        self._content_stream = None
+        self.is_content_rw = False
 
         self.rewrite_types = rewrite_types
 
@@ -287,15 +316,20 @@ class RewriteInfo(object):
         if self.TAG_REGEX.match(buff):
             self.text_type = 'html'
 
+    @property
+    def content_stream(self):
+        if not self._content_stream:
+            self._content_stream = self.record.content_stream()
+            self.is_content_rw = True
+
+        return self._content_stream
+
     def read_and_keep(self, size):
         buff = self.content_stream.read(size)
-        self.content_stream = BufferedReader(self.content_stream, starting_data=buff)
+        self._content_stream = BufferedReader(self._content_stream, starting_data=buff)
         return buff
 
-    def is_content_rw(self):
-        if not self.url_rewriter.prefix:
-            return False
-
+    def should_rw_content(self):
         if self.url_rewriter.wburl.mod == 'id_':
             return False
 
@@ -310,15 +344,15 @@ class RewriteInfo(object):
         elif not self.text_type:
             return False
 
+        elif self.text_type == 'css' or self.text_type == 'xml':
+            if self.url_rewriter.wburl.mod == 'bn_':
+                return False
+
         return True
 
     def is_url_rw(self):
-        if not self.url_rewriter:
-            return False
-
-        if self.url_rewriter.wburl.mod == 'id_':
+        if self.url_rewriter.wburl.mod in ('id_', 'bn_'):
             return False
 
         return True
 
-

pywb/rewrite/html_insert_rewriter.py

@@ -0,0 +1,34 @@
+import re
+from pywb.rewrite.content_rewriter import StreamingRewriter
+
+
+# ============================================================================
+class HTMLInsertOnlyRewriter(StreamingRewriter):
+    """ Insert custom string into HTML <head> tag
+        no other rewriting performed
+    """
+    HEAD_REGEX = re.compile('<\s*head\\b[^>]*[>]+', re.I)
+
+    def __init__(self, url_rewriter, **kwargs):
+        super(HTMLInsertOnlyRewriter, self).__init__(url_rewriter, False)
+        self.head_insert = kwargs['head_insert']
+
+        self.done = False
+
+    def rewrite(self, string):
+        if self.done:
+            return string
+
+        # only try to find <head> in first buffer
+        self.done = True
+        m = self.HEAD_REGEX.search(string)
+        if m:
+            inx = m.end()
+            buff = string[:inx]
+            buff += self.head_insert
+            buff += string[inx:]
+            return buff
+        else:
+            return string
+
+

pywb/rewrite/html_rewriter.py

@@ -19,7 +19,7 @@ from six import text_type
 
 
 #=================================================================
-class HTMLRewriterMixin(object):
+class HTMLRewriterMixin(StreamingRewriter):
     """
     HTML-Parsing Rewriter for custom rewriting, also delegates
     to rewriters for script and css
@@ -98,7 +98,7 @@ class HTMLRewriterMixin(object):
                  defmod='',
                  parse_comments=False):
 
-        self.url_rewriter = url_rewriter
+        super(HTMLRewriterMixin, self).__init__(url_rewriter, False)
         self._wb_parse_context = None
 
         if js_rewriter:
@@ -443,7 +443,7 @@ class HTMLRewriterMixin(object):
 
 
 #=================================================================
-class HTMLRewriter(HTMLRewriterMixin, StreamingRewriter, HTMLParser):
+class HTMLRewriter(HTMLRewriterMixin, HTMLParser):
     PARSETAG = re.compile('[<]')
 
     def __init__(self, *args, **kwargs):
@@ -453,8 +453,6 @@ class HTMLRewriter(HTMLRewriterMixin, StreamingRewriter, HTMLParser):
             HTMLParser.__init__(self)
 
         super(HTMLRewriter, self).__init__(*args, **kwargs)
-        # for StreamingRewriter
-        self.align_to_line = False
 
     def reset(self):
         HTMLParser.reset(self)

pywb/rewrite/jsonp_rewriter.py

@@ -7,10 +7,6 @@ class JSONPRewriter(StreamingRewriter):
     JSONP = re.compile(r'^(\w+)\(\{')
     CALLBACK = re.compile(r'[?].*callback=([^&]+)')
 
-    def __init__(self, urlrewriter):
-        super(JSONPRewriter, self).__init__()
-        self.urlrewriter = urlrewriter
-
     def rewrite(self, string):
         # see if json is jsonp, starts with callback func
         m_json = self.JSONP.search(string)

pywb/rewrite/regex_rewriters.py

@@ -1,6 +1,5 @@
 import re
 
-from pywb.rewrite.url_rewriter import UrlRewriter
 from pywb.rewrite.content_rewriter import StreamingRewriter
 
 
@@ -44,7 +43,7 @@ class RegexRewriter(StreamingRewriter):
     #DEFAULT_OP = add_prefix
 
     def __init__(self, rewriter, rules):
-        super(RegexRewriter, self).__init__()
+        super(RegexRewriter, self).__init__(rewriter)
         #rules = self.create_rules(http_prefix)
 
         # Build regexstr, concatenating regex list

pywb/rewrite/rewrite_amf.py

@@ -1,11 +1,13 @@
 from io import BytesIO
 from six.moves import zip
 
+from pywb.rewrite.content_rewriter import BufferedRewriter
+
 
 # ============================================================================
 # Experimental: not fully tested
-class RewriteAMF(object):  #pragma: no cover
-    def __call__(self, rwinfo):
+class RewriteAMF(BufferedRewriter):  #pragma: no cover
+    def rewrite_stream(self, stream):
         try:
             from pyamf import remoting
 
@@ -20,7 +22,7 @@ class RewriteAMF(object):  #pragma: no cover
             res = remoting.decode(iobuff)
 
             # TODO: revisit this
-            inputdata = rwinfo.url_rewriter.rewrite_opts.get('pywb.inputdata')
+            inputdata = url_rewriter.rewrite_opts.get('pywb.inputdata')
 
             if inputdata:
                 new_list = []
@@ -42,3 +44,5 @@ class RewriteAMF(object):  #pragma: no cover
             traceback.print_exc()
             print(e)
             return stream
+
+

pywb/rewrite/rewrite_dash.py

@@ -4,24 +4,14 @@ import json
 
 import xml.etree.ElementTree as ET
 
-from pywb.webagg.utils import StreamIter
+from pywb.rewrite.content_rewriter import BufferedRewriter
 
 
 # ============================================================================
-class RewriteDASH(object):
-    def __call__(self, rwinfo):
-        buff_io = BytesIO()
-        with closing(rwinfo.content_stream) as fh:
-            while True:
-                buff = fh.read()
-                if not buff:
-                    break
-
-                buff_io.write(buff)
-
-        buff_io.seek(0)
-        res_buff, best_ids = self.rewrite_dash(buff_io)
-        return StreamIter(res_buff)
+class RewriteDASH(BufferedRewriter):
+    def rewrite_stream(self, stream):
+        res_buff, best_ids = self.rewrite_dash(stream)
+        return res_buff
 
     def rewrite_dash(self, stream):
         ET.register_namespace('', 'urn:mpeg:dash:schema:mpd:2011')
@@ -70,7 +60,7 @@ def rewrite_fb_dash(string):
     buff = string.encode('utf-8').decode('unicode-escape')
     buff = buff.encode('utf-8')
     io = BytesIO(buff)
-    io, best_ids = RewriteDASHMixin().rewrite_dash(io)
+    io, best_ids = RewriteDASH().rewrite_dash(io)
     string = json.dumps(io.read().decode('utf-8'))
     string = string[1:-1].replace('<', r'\x3C')
 

pywb/rewrite/rewrite_hls.py

@@ -1,16 +1,14 @@
 import re
 from io import BytesIO
-from pywb.webagg.utils import StreamIter
+
+from pywb.rewrite.content_rewriter import BufferedRewriter
 
 
 # ============================================================================
-class RewriteHLS(object):
+class RewriteHLS(BufferedRewriter):
     EXT_INF = re.compile('#EXT-X-STREAM-INF:(?:.*[,])?BANDWIDTH=([\d]+)')
 
-    def __call__(self, rwinfo):
-        return StreamIter(self.rewrite_m3u8(rwinfo.content_stream))
-
-    def rewrite_m3u8(self, stream):
+    def rewrite_stream(self, stream):
         buff = stream.read()
 
         lines = buff.decode('utf-8').split('\n')

pywb/urlrewrite/header_rewriter.py

@@ -17,6 +17,9 @@ class PrefixHeaderRewriter(object):
         'content-location': 'url-rewrite',
         'content-base': 'url-rewrite',
 
+        'transfer-encoding': 'prefix',
+        'connection': 'prefix',
+
         'content-encoding': 'keep-if-no-content-rewrite',
         'content-length': 'content-length',
 
@@ -24,13 +27,16 @@ class PrefixHeaderRewriter(object):
         'cookie': 'cookie',
     }
 
-    default_rule = 'prefix'
-
     def __init__(self, rwinfo, header_prefix='X-Archive-Orig-'):
         self.header_prefix = header_prefix
         self.rwinfo = rwinfo
         self.http_headers = rwinfo.record.http_headers
 
+        if rwinfo.is_url_rw():
+            self.default_rule = 'prefix'
+        else:
+            self.default_rule = 'keep'
+
     def __call__(self):
         new_headers_list = []
         for name, value in self.http_headers.headers:
@@ -54,14 +60,14 @@ class PrefixHeaderRewriter(object):
             return (name, self.rwinfo.url_rewriter.rewrite(value))
 
         elif rule == 'keep-if-no-content-rewrite':
-            if not self.rwinfo.is_content_rw():
+            if not self.rwinfo.is_content_rw:
                 return (name, value)
 
         elif rule == 'content-length':
             if value == '0':
                 return (name, value)
 
-            if not self.rwinfo.is_content_rw():
+            if not self.rwinfo.is_content_rw:
                 try:
                     if int(value) >= 0:
                         return (name, value)
@@ -92,11 +98,3 @@ class PrefixHeaderRewriter(object):
             new_headers.append(('Expires', datetime_to_http_date(dt)))
 
 
-#=============================================================================
-class ProxyHeaderRewriter(PrefixHeaderRewriter):
-    header_rules = {
-        'transfer-encoding': 'prefix',
-        'connection': 'prefix',
-    }
-
-    default_rule = 'keep'

pywb/urlrewrite/rewriter.py

@@ -1,12 +1,13 @@
 from pywb.rewrite.content_rewriter import BaseContentRewriter
 
 from pywb.rewrite.html_rewriter import HTMLRewriter
+from pywb.rewrite.html_insert_rewriter import HTMLInsertOnlyRewriter
 
 from pywb.rewrite.regex_rewriters import RegexRewriter, CSSRewriter, XMLRewriter
 from pywb.rewrite.regex_rewriters import JSLinkAndLocationRewriter, JSLinkOnlyRewriter
 from pywb.rewrite.regex_rewriters import JSLocationOnlyRewriter, JSNoneRewriter
 
-from pywb.urlrewrite.header_rewriter import PrefixHeaderRewriter, ProxyHeaderRewriter
+from pywb.urlrewrite.header_rewriter import PrefixHeaderRewriter
 
 from pywb.rewrite.jsonp_rewriter import JSONPRewriter
 
@@ -19,9 +20,9 @@ from pywb.rewrite.rewrite_amf import RewriteAMF
 class DefaultRewriter(BaseContentRewriter):
     all_rewriters = {
         'header': PrefixHeaderRewriter,
-        'header-proxy': ProxyHeaderRewriter,
 
         'html': HTMLRewriter,
+        'html-banner-only': HTMLInsertOnlyRewriter,
 
         'css': CSSRewriter,
 

tests/test_integration.py

@@ -152,7 +152,7 @@ class TestWbIntegration(BaseConfigTest):
         assert len(lines) == 17
         assert lines[0].startswith('org,iana)/_css/2013.1/print.css 20140127171239')
 
-    def _test_replay_banner_only(self):
+    def test_replay_banner_only(self):
         resp = self.testapp.get('/pywb/20140126201054bn_/http://www.iana.org/domains/reserved')
 
         # wb.js header insertion