11# -*- coding: utf-8 -*-
22"""
3- :copyright: Copyright 2018 by the contributors (see AUTHORS file).
3+ :copyright: Copyright 2018-2019 by the contributors (see AUTHORS file).
44 :license: BSD-2-Clause, see LICENSE for details.
55"""
66
@@ -58,8 +58,9 @@ def __init__(self, master, env):
5858 self .assets = []
5959 self .env = env
6060 self .hash2asset = {}
61- self .key2asset = {}
61+ self .keys = set ()
6262 self .master = master
63+ self .path2asset = {}
6364
6465 def build (self ):
6566 """
@@ -88,9 +89,9 @@ def build(self):
8889 data .append (entry )
8990 return data
9091
91- def asset2docname (self , key ):
92+ def fetch (self , node ):
9293 """
93- return target document name for provided asset
94+ return key and target document name for provided asset
9495
9596 When given a asset, cached information will return the name of the
9697 target document this asset will be published to. In the event an asset
@@ -99,20 +100,26 @@ def asset2docname(self, key):
99100 document name will be returned instead.
100101
101102 Args:
102- key : the asset key
103+ node : the node to interpret
103104
104105 Returns:
105- the document name
106+ the key and document name
106107 """
107108 docname = None
108- asset = self .key2asset .get (key , None )
109- if asset :
110- if len (asset .docnames ) > 1 :
111- docname = self .master
112- else :
113- docname = next (iter (asset .docnames ))
109+ key = None
114110
115- return docname
111+ path = self ._interpretAssetPath (node )
112+ if path :
113+ asset = self .path2asset .get (path , None )
114+ if asset :
115+ key = asset .key
116+
117+ if len (asset .docnames ) > 1 :
118+ docname = self .master
119+ else :
120+ docname = next (iter (asset .docnames ))
121+
122+ return key , docname
116123
117124 def process (self , docnames ):
118125 """
@@ -132,7 +139,7 @@ def process(self, docnames):
132139
133140 def processDocument (self , doctree , docname , standalone = False ):
134141 """
135- process a docment for assets
142+ process a document for assets
136143
137144 This method will search each the provided document's doctree for
138145 supported assets which could be published. Asset information is tracked
@@ -148,117 +155,111 @@ def processDocument(self, doctree, docname, standalone=False):
148155 for node in image_nodes :
149156 uri = node ['uri' ]
150157 if not uri .startswith ('data:' ) and uri .find ('://' ) == - 1 :
151- key , path = self .interpretAssetKeyPath (node )
152- if key not in self .key2asset :
158+ path = self ._interpretAssetPath (node )
159+ if path not in self .path2asset :
153160 hash = ConfluenceUtil .hashAsset (path )
154161 type = guess_mimetype (path , default = DEFAULT_CONTENT_TYPE )
155162 else :
156- hash = self .key2asset [ key ].hash
157- type = self .key2asset [ key ].type
158- self ._handleEntry (key , path , type , hash , docname , standalone )
163+ hash = self .path2asset [ path ].hash
164+ type = self .path2asset [ path ].type
165+ self ._handleEntry (path , type , hash , docname , standalone )
159166
160167 file_nodes = doctree .traverse (addnodes .download_reference )
161168 for node in file_nodes :
162169 target = node ['reftarget' ]
163170 if target .find ('://' ) == - 1 :
164- key , path = self .interpretAssetKeyPath (node )
165- if key not in self .key2asset :
171+ path = self ._interpretAssetPath (node )
172+ if path not in self .path2asset :
166173 hash = ConfluenceUtil .hashAsset (path )
167174 type = guess_mimetype (path , default = DEFAULT_CONTENT_TYPE )
168175 else :
169- hash = self .key2asset [ key ].hash
170- type = self .key2asset [ key ].type
171- self ._handleEntry (key , path , type , hash , docname , standalone )
176+ hash = self .path2asset [ path ].hash
177+ type = self .path2asset [ path ].type
178+ self ._handleEntry (path , type , hash , docname , standalone )
172179
173- def _handleEntry (self , key , path , type , hash , docname , standalone = False ):
180+ def _handleEntry (self , path , type , hash , docname , standalone = False ):
174181 """
175182 handle an asset entry
176183
177184 When an asset is detected in a document, the information about the asset
178185 is tracked in this manager. When an asset is detected, there are
179- considerations to be made. If an asset key has already been registered
180- (e.x . an asset used twice), only a single asset entry will be created.
186+ considerations to be made. If an asset path has already been registered
187+ (e.g . an asset used twice), only a single asset entry will be created.
181188 If an asset matches the hash of another asset, another entry is *not
182189 created (i.e. a documentation set has duplicate assets; *with the
183190 exception of when ``standalone`` is set to ``True``). In all cases where
184191 an asset is detected, the asset reference is updated to track which
185192 document the asset belongs to.
186193
187194 Args:
188- key: the asset key
189195 path: the absolute path to the asset
190196 type: the content type of the asset
191197 hash: the hash of the asset
192198 docname: the document name this asset was found in
193199 standalone (optional): ignore hash mappings (defaults to False)
194200 """
195- asset = self .key2asset .get (key , None )
201+ asset = self .path2asset .get (path , None )
196202 if not asset :
197203 hash_exists = hash in self .hash2asset
198204 if not hash_exists or standalone :
199205 # no asset entry and no hash entry (or standalone); new asset
206+ key = os .path .basename (path )
207+
208+ # Confluence does not allow attachments with select characters.
209+ # Filter out the asset name to a compatible key value.
210+ for rep in INVALID_CHARS :
211+ key = key .replace (rep , '_' )
212+
213+ filename , file_ext = os .path .splitext (key )
214+ idx = 1
215+ while key in self .keys :
216+ idx += 1
217+ key = '{}_{}{}' .format (filename , idx , file_ext )
218+ self .keys .add (key )
219+
200220 asset = ConfluenceAsset (key , path , type , hash )
201221 self .assets .append (asset )
202- self .key2asset [ key ] = asset
222+ self .path2asset [ path ] = asset
203223 if not hash_exists :
204- self .hash2asset [key ] = asset
224+ self .hash2asset [hash ] = asset
205225 else :
206226 # duplicate asset detected; build an asset alias
207227 asset = self .hash2asset [hash ]
208- self .key2asset [ key ] = asset
228+ self .path2asset [ path ] = asset
209229 else :
210- assert (self .hash2asset [key ] == asset )
230+ assert (self .hash2asset [asset . hash ] == asset )
211231
212232 # track (if not already) that this document uses this asset
213233 asset .docnames .add (docname )
214234
215- def interpretAssetKeyPath (self , node ):
235+ def _interpretAssetPath (self , node ):
216236 """
217- find a key value and absolute path for a target assert
237+ find an absolute path for a target assert
218238
219- Returns a "key" value (a normalized path to the asset relative to the
220- documentation's root) as well as the absolute path to the assert. For
221- unsupported asset types, this method will return ``None`` values. This
222- method should not be invoked on external assets (i.e. URLs).
239+ Returns the absolute path to an assert. For unsupported asset types,
240+ this method will return ``None`` values. This method should not be
241+ invoked on external assets (i.e. URLs).
223242
224243 Args:
225244 node: the node to parse
226245
227246 Returns:
228- the key and absolute path
247+ the absolute path
229248 """
230- key = None
249+ path = None
231250 if isinstance (node , nodes .image ):
232- # uri's will be relative to documentation root. Normalize for a key
233- # value to handle assets found in parent directories.
234- uri = node ['uri' ]
235- uri = os .path .normpath (uri )
236- path = uri
237-
238- # If this URI is found outside the relative path of the
239- # documentation set, grab the basename and add a prefix for
240- # (hopefully) a unique name.
241- if os .path .isabs (uri ):
242- key = 'extern_{}' .format (os .path .basename (uri ))
243- else :
244- key = uri
251+ # uri's will be relative to documentation root.
252+ path = node ['uri' ]
245253 elif isinstance (node , addnodes .download_reference ):
246254 # reftarget will be a reference to the asset with respect to the
247255 # document (refdoc) holding this reference. Use reftarget and refdoc
248- # to find a proper key value (normalized to handle parent directory
249- # references).
256+ # to find a proper path.
250257 docdir = os .path .dirname (node ['refdoc' ])
251- key = os .path .join (docdir , node ['reftarget' ])
252- key = os .path .normpath (key )
253- path = key
258+ path = os .path .join (docdir , node ['reftarget' ])
254259
255260 abspath = None
256- if key :
261+ if path :
262+ path = os .path .normpath (path )
257263 abspath = os .path .join (self .env .srcdir , path )
258264
259- # Confluence does not allow attachments with select characters.
260- # Filter out the asset name to a compatible key value.
261- for rep in INVALID_CHARS :
262- key = key .replace (rep , '_' )
263-
264- return key , abspath
265+ return abspath
0 commit comments