def rev(tree, revision, path):
"""Display a page showing the file at path at specified revision by
obtaining the contents from version control.
"""
config = current_app.dxr_config
tree_config = config.trees[tree]
abs_path = join(tree_config.source_folder, path)
contents = file_contents_at_rev(abs_path, revision)
if contents is not None:
image_rev = None
if is_binary_image(path):
is_text = False
contents = ''
image_rev = revision
else:
is_text, contents = decode_data(contents, tree_config.source_encoding)
if not is_text:
contents = ''
elif is_textual_image(path):
image_rev = revision
# We do some wrapping to mimic the JSON returned by an ES lines query.
return _browse_file(tree,
path,
[{'content': line} for line in split_content_lines(contents)],
{},
config,
not is_text,
contents=contents,
image_rev=image_rev)
else:
raise NotFound
def raw_rev(tree, revision, path):
"""Send raw data at path from tree at the given revision, for binary things
like images."""
if not is_binary_image(path) and not is_textual_image(path):
raise NotFound
config = current_app.dxr_config
tree_config = config.trees[tree]
data = file_contents_at_rev(tree_config.source_folder, path, revision)
if data is None:
raise NotFound
data_file = StringIO(data)
return send_file(data_file, mimetype=guess_type(path)[0])
def raw(tree, path):
"""Send raw data at path from tree, for binary things like images."""
if not is_binary_image(path) and not is_textual_image(path):
raise NotFound
query = {'filter': {'term': {'path': path}}}
results = current_app.es.search(query,
index=es_alias_or_not_found(tree),
doc_type=FILE,
size=1)
try:
# we explicitly get index 0 because there should be exactly 1 result
data = results['hits']['hits'][0]['_source']['raw_data'][0]
except IndexError: # couldn't find the image
raise NotFound
data_file = StringIO(data.decode('base64'))
return send_file(data_file, mimetype=guess_type(path)[0])
def needles(self):
"""Fill out path (and path.trigrams)."""
if self.is_link():
# realpath will keep following symlinks until it gets to the 'real' thing.
yield 'link', relpath(realpath(self.absolute_path()), self.tree.source_folder)
yield 'path', self.path
extension = splitext(self.path)[1]
if extension:
yield 'ext', extension[1:] # skip the period
# We store both the contents of textual images twice so that they can
# both show up in searches and be previewed in the browser.
if is_binary_image(self.path) or is_textual_image(self.path):
bytestring = (self.contents.encode('utf-8') if self.contains_text()
else self.contents)
yield 'raw_data', b64encode(bytestring)
# binary, but not an image
elif not self.contains_text():
yield 'is_binary', True
def needles(self):
"""Fill out path (and path.trigrams)."""
if self.is_link():
# realpath will keep following symlinks until it gets to the 'real' thing.
yield 'link', relpath(realpath(self.absolute_path()),
self.tree.source_folder)
yield 'path', self.path
extension = splitext(self.path)[1]
if extension:
yield 'ext', extension[1:] # skip the period
# We store both the contents of textual images twice so that they can
# both show up in searches and be previewed in the browser.
if is_binary_image(self.path) or is_textual_image(self.path):
bytestring = (self.contents.encode('utf-8')
if self.contains_text() else self.contents)
yield 'raw_data', b64encode(bytestring)
# binary, but not an image
elif not self.contains_text():
yield 'is_binary', True
def needles(self):
"""Fill out path (and path.trigrams)."""
if self.is_link():
# realpath will keep following symlinks until it gets to the 'real' thing.
yield 'link', relpath(realpath(self.absolute_path()),
self.tree.source_folder)
unicode_path = unicode_for_display(self.path)
yield 'path', unicode_path
yield 'file_name', basename(unicode_path)
extension = splitext(unicode_path)[1]
if extension:
yield 'ext', extension[1:] # skip the period
# We store both the contents of textual images twice so that they can
# both show up in searches and be previewed in the browser.
if is_binary_image(self.path) or is_textual_image(self.path):
# If the file was binary, then contents are None, so read it here.
if self.contents is None:
with open(self.absolute_path(), 'rb') as image_file:
self.contents = image_file.read()
bytestring = (self.contents.encode('utf-8') if self.contains_text()
else self.contents)
yield 'raw_data', b64encode(bytestring)
# binary, but not an image
elif not self.contains_text():
yield 'is_binary', True
# Find the last modified time from version control if possible,
# otherwise fall back to the timestamp from stat'ing the file.
modified = None
if self.vcs:
vcs_relative_path = relpath(self.absolute_path(),
self.vcs.get_root_dir())
try:
modified = self.vcs.last_modified_date(vcs_relative_path)
except NotImplementedError:
pass
if modified is None:
file_info = stat(self.absolute_path())
modified = datetime.utcfromtimestamp(file_info.st_mtime)
yield 'modified', modified
def raw(tree, path):
"""Send raw data at path from tree, for binary things like images."""
if not is_binary_image(path) and not is_textual_image(path):
raise NotFound
query = {
'filter': {
'term': {
'path': path
}
}
}
results = current_app.es.search(
query,
index=es_alias_or_not_found(tree),
doc_type=FILE,
size=1)
try:
# we explicitly get index 0 because there should be exactly 1 result
data = results['hits']['hits'][0]['_source']['raw_data'][0]
except IndexError: # couldn't find the image
raise NotFound
data_file = StringIO(data.decode('base64'))
return send_file(data_file, mimetype=guess_type(path)[0])
def _browse_file(tree, path, line_docs, file_doc, config, is_binary,
date=None, contents=None, image_rev=None):
"""Return a rendered page displaying a source file.
:arg string tree: name of tree on which file is found
:arg string path: relative path from tree root of file
:arg list line_docs: LINE documents as defined in the mapping of core.py,
where the `content` field is dereferenced
:arg file_doc: the FILE document as defined in core.py
:arg config: TreeConfig object of this tree
:arg is_binary: Whether file is binary or not
:arg date: a formatted string representing the generated date, default to now
:arg string contents: the contents of the source file, defaults to joining
the `content` field of all line_docs
:arg image_rev: revision number of a textual or binary image, for images
displayed at a certain rev
"""
def process_link_templates(sections):
"""Look for {{line}} in the links of given sections, and duplicate them onto
a 'template' field.
"""
for section in sections:
for link in section['items']:
if '{{line}}' in link['href']:
link['template'] = link['href']
link['href'] = link['href'].replace('{{line}}', '')
def sidebar_links(sections):
"""Return data structure to build nav sidebar from. ::
[('Section Name', [{'icon': ..., 'title': ..., 'href': ...}])]
"""
process_link_templates(sections)
# Sort by order, resolving ties by section name:
return sorted(sections, key=lambda section: (section['order'],
section['heading']))
if not date:
# Then assume that the file is generated now. Remark: we can't use this
# as the default param because that is only evaluated once, so the same
# time would always be used.
date = datetime.utcnow().strftime("%a, %d %b %Y %H:%M:%S +0000")
common = _build_common_file_template(tree, path, is_binary, date, config)
links = file_doc.get('links', [])
if is_binary_image(path):
return render_template(
'image_file.html',
**merge(common, {
'sections': sidebar_links(links),
'revision': image_rev}))
elif is_binary:
return render_template(
'text_file.html',
**merge(common, {
'lines': [],
'is_binary': True,
'sections': sidebar_links(links)}))
else:
# We concretize the lines into a list because we iterate over it multiple times
lines = [doc['content'] for doc in line_docs]
if not contents:
# If contents are not provided, we can reconstruct them by
# stitching the lines together.
contents = ''.join(lines)
offsets = build_offset_map(lines)
tree_config = config.trees[tree]
if is_textual_image(path) and image_rev:
# Add a link to view textual images on revs:
links.extend(dictify_links([
(4,
'Image',
[('svgview', 'View', url_for('.raw_rev',
tree=tree_config.name,
path=path,
revision=image_rev))])]))
# Construct skimmer objects for all enabled plugins that define a
# file_to_skim class.
skimmers = [plugin.file_to_skim(path,
contents,
plugin.name,
tree_config,
file_doc,
line_docs)
for plugin in tree_config.enabled_plugins
if plugin.file_to_skim]
skim_links, refses, regionses, annotationses = skim_file(skimmers, len(line_docs))
index_refs = (Ref.es_to_triple(ref, tree_config) for ref in
chain.from_iterable(doc.get('refs', [])
for doc in line_docs))
index_regions = (Region.es_to_triple(region) for region in
chain.from_iterable(doc.get('regions', [])
for doc in line_docs))
tags = finished_tags(lines,
chain(chain.from_iterable(refses), index_refs),
chain(chain.from_iterable(regionses), index_regions))
return render_template(
'text_file.html',
**merge(common, {
# Someday, it would be great to stream this and not concretize
# the whole thing in RAM. The template will have to quit
# looping through the whole thing 3 times.
'lines': [(html_line(doc['content'], tags_in_line, offset),
doc.get('annotations', []) + skim_annotations)
for doc, tags_in_line, offset, skim_annotations
in izip(line_docs, tags_per_line(tags), offsets, annotationses)],
'sections': sidebar_links(links + skim_links),
'query': request.args.get('q', ''),
'bubble': request.args.get('redirect_type')}))
def _browse_file(tree,
path,
line_docs,
file_doc,
config,
date=None,
contents=None):
"""Return a rendered page displaying a source file.
:arg string tree: name of tree on which file is found
:arg string path: relative path from tree root of file
:arg list line_docs: LINE documents as defined in the mapping of core.py,
where the `content` field is dereferenced
:arg file_doc: the FILE document as defined in core.py
:arg config: TreeConfig object of this tree
:arg date: a formatted string representing the generated date, default to now
:arg string contents: the contents of the source file, defaults to joining
the `content` field of all line_docs
"""
def sidebar_links(sections):
"""Return data structure to build nav sidebar from. ::
[('Section Name', [{'icon': ..., 'title': ..., 'href': ...}])]
"""
# Sort by order, resolving ties by section name:
return sorted(sections,
key=lambda section:
(section['order'], section['heading']))
if not date:
# Then assume that the file is generated now. Remark: we can't use this
# as the default param because that is only evaluated once, so the same
# time would always be used.
date = datetime.utcnow().strftime("%a, %d %b %Y %H:%M:%S +0000")
common = _build_common_file_template(tree, path, date, config)
links = file_doc.get('links', [])
if is_binary_image(path):
return render_template('image_file.html', **common)
else: # We don't allow browsing binary files, so this must be a text file.
# We concretize the lines into a list because we iterate over it multiple times
lines = [doc['content'] for doc in line_docs]
if not contents:
# If contents are not provided, we can reconstruct them by
# stitching the lines together.
contents = ''.join(lines)
offsets = cumulative_sum(imap(len, lines))
tree_config = config.trees[tree]
# Construct skimmer objects for all enabled plugins that define a
# file_to_skim class.
skimmers = [
plugin.file_to_skim(path, contents, plugin.name, tree_config,
file_doc, line_docs)
for plugin in tree_config.enabled_plugins if plugin.file_to_skim
]
skim_links, refses, regionses, annotationses = skim_file(
skimmers, len(line_docs))
index_refs = (Ref.es_to_triple(ref, tree_config)
for ref in chain.from_iterable(
doc.get('refs', []) for doc in line_docs))
index_regions = (Region.es_to_triple(region)
for region in chain.from_iterable(
doc.get('regions', []) for doc in line_docs))
tags = finished_tags(
lines, chain(chain.from_iterable(refses), index_refs),
chain(chain.from_iterable(regionses), index_regions))
return render_template(
'text_file.html',
**merge(
common,
{
# Someday, it would be great to stream this and not concretize
# the whole thing in RAM. The template will have to quit
# looping through the whole thing 3 times.
'lines': [(html_line(doc['content'], tags_in_line, offset),
doc.get('annotations', []) + skim_annotations)
for doc, tags_in_line, offset, skim_annotations
in izip(line_docs, tags_per_line(tags), offsets,
annotationses)],
'is_text':
True,
'sections':
sidebar_links(links + skim_links)
}))
def _browse_file(tree, path, line_docs, file_doc, config, date=None, contents=None):
"""Return a rendered page displaying a source file.
:arg string tree: name of tree on which file is found
:arg string path: relative path from tree root of file
:arg list line_docs: LINE documents as defined in the mapping of core.py,
where the `content` field is dereferenced
:arg file_doc: the FILE document as defined in core.py
:arg config: TreeConfig object of this tree
:arg date: a formatted string representing the generated date, default to now
:arg string contents: the contents of the source file, defaults to joining
the `content` field of all line_docs
"""
def sidebar_links(sections):
"""Return data structure to build nav sidebar from. ::
[('Section Name', [{'icon': ..., 'title': ..., 'href': ...}])]
"""
# Sort by order, resolving ties by section name:
return sorted(sections, key=lambda section: (section['order'],
section['heading']))
if not date:
# Then assume that the file is generated now. Remark: we can't use this
# as the default param because that is only evaluated once, so the same
# time would always be used.
date = datetime.utcnow().strftime("%a, %d %b %Y %H:%M:%S +0000")
common = _build_common_file_template(tree, path, date, config)
links = file_doc.get('links', [])
if is_binary_image(path):
return render_template(
'image_file.html',
**common)
else: # We don't allow browsing binary files, so this must be a text file.
# We concretize the lines into a list because we iterate over it multiple times
lines = [doc['content'] for doc in line_docs]
if not contents:
# If contents are not provided, we can reconstruct them by
# stitching the lines together.
contents = ''.join(lines)
offsets = cumulative_sum(imap(len, lines))
tree_config = config.trees[tree]
# Construct skimmer objects for all enabled plugins that define a
# file_to_skim class.
skimmers = [plugin.file_to_skim(path,
contents,
plugin.name,
tree_config,
file_doc,
line_docs)
for plugin in tree_config.enabled_plugins
if plugin.file_to_skim]
skim_links, refses, regionses, annotationses = skim_file(skimmers, len(line_docs))
index_refs = (Ref.es_to_triple(ref, tree_config) for ref in
chain.from_iterable(doc.get('refs', [])
for doc in line_docs))
index_regions = (Region.es_to_triple(region) for region in
chain.from_iterable(doc.get('regions', [])
for doc in line_docs))
tags = finished_tags(lines,
chain(chain.from_iterable(refses), index_refs),
chain(chain.from_iterable(regionses), index_regions))
return render_template(
'text_file.html',
**merge(common, {
# Someday, it would be great to stream this and not concretize
# the whole thing in RAM. The template will have to quit
# looping through the whole thing 3 times.
'lines': [(html_line(doc['content'], tags_in_line, offset),
doc.get('annotations', []) + skim_annotations)
for doc, tags_in_line, offset, skim_annotations
in izip(line_docs, tags_per_line(tags), offsets, annotationses)],
'is_text': True,
'sections': sidebar_links(links + skim_links)}))