Class: LinkHeaders::Processor

Inherits:
Object
  • Object
show all
Defined in:
lib/linkheaders/processor.rb,
lib/linkheaders/processor/version.rb

Overview

A Link Header parser

Works for both HTML and HTTP links, and handles references to Linksets of either JSON or Text types

Constant Summary collapse

VERSION = 
"0.1.21"

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(default_anchor: 'https://default.anchor.org/') ⇒ Processor

Create the Link Headers Parser and its Link factory

Parameters:

  • default_anchor (<String>) (defaults to: 'https://default.anchor.org/') —

    Link relations always have an anchor, but it is sometimes implicit. This value will be used in implicit cases.



28
29
30
31
 
# File 'lib/linkheaders/processor.rb', line 28

def initialize(default_anchor: 'https://default.anchor.org/')
  @default_anchor = default_anchor
  @factory = LinkHeaders::LinkFactory.new(default_anchor: @default_anchor)
end

Instance Attribute Details

#default_anchor ⇒ <Type>

Returns <description>.

Returns:

  • (<Type>) —

    <description>



21
22
23
 
# File 'lib/linkheaders/processor.rb', line 21

def default_anchor
  @default_anchor
end

#factory ⇒ <LinkHeader::LinkFactory>

Get the parser factory that contains all the links

Returns:

  • (<LinkHeader::LinkFactory>) —

    The factory containing the links (LinkHeader::Link) that have been created so far



21
22
23
 
# File 'lib/linkheaders/processor.rb', line 21

def factory
  @factory
end

Instance Method Details

#check_for_linkset(responsepart:) ⇒ Object 



175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
 
# File 'lib/linkheaders/processor.rb', line 175

def check_for_linkset(responsepart:)
  warn "looking for a linkset"
  newlinks = Array.new
  factory.linksets.each do |linkset|
    # warn "found #{linkset.methods- Object.new.methods}"
    # warn "inspect #{linkset.inspect}"
    next unless linkset.respond_to? 'type'
    # warn "responds #{linkset.type}  "
    case linkset.type
    when 'application/linkset+json'
      # warn "found a json linkset"
      newlinks << processJSONLinkset(href: linkset.href)
    when 'application/linkset'
      # warn "found a text linkset"
      newlinks << processTextLinkset(href:linkset.href)
    else
      warn "the linkset #{linkset} was not typed as 'application/linkset+json' or 'application/linkset', and it should be! (found #{linkset.type}) Ignoring..."
    end
  end
  newlinks
end

#extract_and_parse(response: RestClient::Response.new) ⇒ Object

Parses a RestClient::Response

The HTTP headers are parsed for Links and if those links contain a Linkset, that is retrieved and parsed If the Response is of some HTML form, this is also parsed for Link headers and Linkset links All discovered links end up in a LinkHeader::LinkFactory object (self.factory)

Parameters:

  • response (<RestClilent::Response>) (defaults to: RestClient::Response.new) —

    The full response object from an HTTP 2** successful call



52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
 
# File 'lib/linkheaders/processor.rb', line 52

def extract_and_parse(response: RestClient::Response.new)
  head = response.headers
  body = response.body
  # warn "\n\n head #{head.inspect}\n\n"

  unless head
    warn "WARNING: This doesn't seem to be a RestClient response message.\nReturning blank"
    return [[], []]
  end

  _newlinks = parse_http_link_headers(head)
  # warn "HTTPlinks #{newlinks.inspect}"

  ['text/html', 'text/xhtml+xml', 'application/xhtml+xml'].each do |format|
    if head[:content_type] and head[:content_type].match(format)
      warn "found #{format} content - parsing"
      _htmllinks = parse_html_link_headers(body: body, anchor: default_anchor) # pass html body to find HTML link headers
      # warn "htmllinks #{htmllinks.inspect}"
    end
  end
end

#parse_html_link_headers(body:, anchor: '') ⇒ Object

Parses the link headers out of an HTML body, and adds links to the LinkHeader::LinkFactory object. Will automatically retrieve and process any LinkSet references found

Parameters:

  • body (String) —

    The HTML of the page containing HTML Link headers



149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
 
# File 'lib/linkheaders/processor.rb', line 149

def parse_html_link_headers(body:, anchor: '')
  m = MetaInspector.new(anchor, document: body)
  # an array of elements that look like this: [{:rel=>"alternate", :type=>"application/ld+json", :href=>"http://scidata.vitk.lv/dataset/303.jsonld"}]
  newlinks = Array.new
  m.head_links.each do |l|
    next unless l[:href] and l[:rel] # required

    anchor = l[:anchor] || default_anchor
    l.delete(:anchor)
    relation = l[:rel]
    l.delete(:rel)
    href = l[:href]
    l.delete(:href)        
    
    relations = relation.split(/\s+/)  # handle the multiple relation case
    # warn "BODY RELATIONS #{relations}"

    relations.each do |rel|
      next unless rel.match?(/\w/)
      newlinks << factory.new_link(responsepart: :header, anchor: anchor, href: href, relation: rel, **l) # parsed['https://example.one.com'][:rel] = "preconnect"
    end
  end
  newlinks << check_for_linkset(responsepart: :body)
  newlinks
end

#parse_http_link_headers(headers) ⇒ Object

Consume a String of the Link Headers and parse it into individual links. Will automatically retrieve and process any LinkSet references found. All LinkHeader::Link objects end up in the LinkHeader::LinkFactory object (self.factory)

Parameters:



80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
 
# File 'lib/linkheaders/processor.rb', line 80

def parse_http_link_headers(headers)
  newlinks = Array.new
  # Link: <https://example.one.com>; rel="preconnect", <https://example.two.com>; rel="preconnect",  <https://example.three.com>; rel="preconnect"
  links = headers[:link]
  return [] unless links

  # warn links.inspect
  parts = links.split(',') # ["<https://example.one.com>; rel='preconnect'", "<https://example.two.com>; rel="preconnect"".....]
  # warn parts

  # Parse each part into a named link
  newlinks << split_http_link_headers_and_process(parts) # creates links from the split headers and adds to factory.all_links
  newlinks << check_for_linkset(responsepart: :header)  # all links are held in the Linkset::LinkFactory object (factory variable here).  This scans the links for a linkset link to follow
  newlinks
end

#processJSONLinkset(href:) ⇒ Object 



197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
 
# File 'lib/linkheaders/processor.rb', line 197

def processJSONLinkset(href:)
  _headers, linkset = lhfetch(href, { 'Accept' => 'application/linkset+json' })
  # warn "Linkset body #{linkset.inspect}\n\nLinkset headers #{_headers}\n\n"
  newlinks = Array.new
  return nil unless linkset
  # warn "linkset #{linkset}"

  # linkset = '{ "linkset":
  #   [
  #     { "anchor": "http://example.net/bar",
  #       "item": [
  #         {"href": "http://example.com/foo1", "type": "text/html"},
  #         {"href": "http://example.com/foo2"}
  #       ],
  #       "next": [
  #         {"href": "http://the.next/"}
  #       ]
  #     }
  #   ]
  # }'

  linkset = JSON.parse(linkset)
  # warn "linkset #{linkset}"
  if linkset['data'] and linkset['data']['linkset']
    linkset['linkset'] = linkset['data']['linkset']
  end
  return nil unless linkset['linkset'].first
  linkset['linkset'].each do |ls|
    # warn ls.inspect, "\n"
    anchor = ls['anchor'] || @default_anchor
    ls.delete('anchor') if ls['anchor'] # we need to delete since almost all others have a list as a value
    attrhash = {}
    # warn ls.keys, "\n"

    ls.each_key do |relation| # relation =  e.g. "item", "described-by". "cite"
      href = ""
      # warn relation
      # warn ls[reltype], "\n"
      ls[relation] = [ls[relation]] unless ls[relation].is_a? Array  # force it to be a list, if it isn't
      ls[relation].each do |attrs|  # attr = e.g.  {"href": "http://example.com/foo1", "type": "text/html"}
        # warn "ATTR: #{attrs}"
        next unless attrs['href']  # this is a required attribute of a  linkset relation
        href = attrs['href']
        attrs.delete("href")
        # now go through the other attributes of that relation
        attrs.each do |attr, val| # attr = e.g. "type"; val = "text/html"
          attrhash[attr.to_sym] = val
        end
      end

      relations = relation.split(/\s+/)  # handle the multiple relation case
      relations.each do |rel|
        next unless rel.match?(/\w/)
        newlinks << factory.new_link(responsepart: :header, anchor: anchor, href: href, relation: rel, **attrhash) # parsed['https://example.one.com'][:rel] = "preconnect"
      end
    end
  end
  newlinks
end

#processTextLinkset(href:) ⇒ Object 



257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
 
# File 'lib/linkheaders/processor.rb', line 257

def processTextLinkset(href:)
  newlinks = Array.new
  headers, linkset = lhfetch(href, { 'Accept' => 'application/linkset' })
  # warn "linkset body #{linkset.inspect}"
  return {} unless linkset

#      links = linkset.scan(/(<.*?>[^<]+)/) # split on the open angle bracket, which indicates a new link
  links = linkset.split(/,\n*/) # split on the comma+newline
  # warn "Links found #{links}"

  links.each do |ls|
    # warn "working on link #{ls}"
    elements = ls.split(';').map {|element| element.strip!} # semicolon delimited fields
    # ["<https://w3id.org/a2a-fair-metrics/08-http-describedby-citeas-linkset-txt/>", "anchor=\"https://s11.no/2022/a2a-fair-metrics/08-http-describedby-citeas-linkset-txt/\"", "rel=\"cite-as\""]
    href = elements.shift # first element is always the link url
    # warn "working on link href #{href}"
    href = href.match(/<([^>]+)>/)[1]
    attrhash = {}
    elements.each do |e|
      key, val = e.split('=')
      val.delete_prefix!('"').delete_suffix!('"') # get rid of newlines and start/end quotes
      attrhash[key.to_sym] = val # split on key=val and make key a symbol
    end 
    warn "No link relation type... this is bad!  Skipping" unless attrhash[:rel]
    next unless attrhash[:rel]
    relation = attrhash[:rel]
    attrhash.delete(:rel)
    anchor = attrhash[:anchor] || @default_anchor
    attrhash.delete(:anchor)
    
    relations = relation.split(/\s+/)  # handle the multiple relation case
    #$stderr.puts "RELATIONS #{relations}"
    relations.each do |rel|
      next unless rel.match?(/\w/)
      newlinks << factory.new_link(responsepart: :header, anchor: anchor, href: href, relation: rel, **attrhash) # parsed['https://example.one.com'][:rel] = "preconnect"
    end
  end
  newlinks
end

#split_http_link_headers_and_process(parts) ⇒ Object 



96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
 
# File 'lib/linkheaders/processor.rb', line 96

def split_http_link_headers_and_process(parts)
  newlinks = Array.new
  parts.each do |part, _index|
    warn "link is:  #{part}"  # <https://s11.no/2022/a2a-fair-metrics/70-rda-r1-01m-t2-type-charset/test-apple-data.csv>;rel="item";type="text/csv;charset=UTF-8"

    # this is crazy hard, because we can't rely on quotes!
    part = part + ";"
    href = part[/<([^\]]+)>;/, 1]
    next unless href
    part = part.gsub(/<([^\]]+)>\s?;\s?/, "") # now only ;rel="item";type="text/csv;charset=UTF-8"; pro=gram
    pieces = part.scan(/(\S+?\s?=\s?"[^"]+?"\s?);?/).flatten  # ["rel=\"item\"", "type=\"text/csv;charset=UTF-8\""]   and the non-quoted stuff is ignored
    pieces.each do |p|
      part = part.gsub(p, "")  # now just ";;prog=gram;"
    end
    rest = part.split(";")  # ["", "", "prog=gram"]

    sections = {}
    pieces.concat(rest).each do |s| # can be more than one link property "rel='preconnect'"
      s.strip!
      unless m = s.match(%r{(\w+?)\s?=\s?"?([\w\;\:\d\.\,\=\#\-\+\/\s]+)"?})
      warn " NO PATTERN MATCH ON #{s}"
        next
      end # can be rel="cite-as describedby"  --> two relations in one!  or "linkset+json"

      relation = m[1] # rel"
      value = m[2] # "preconnect"
      warn "section relation #{relation} value #{value}"
      sections[relation] = value # value could hold multiple relation types   sections[:rel] = "preconnect"
    end

    next unless sections['rel']  # the relation is required!

    anchor = sections['anchor'] || default_anchor
    sections.delete('anchor')
    relation = sections['rel']
    sections.delete('rel')
    relations = relation.split(/\s+/)  # handle the multiple relation case
    # warn "HEADERS RELATIONS #{relations}"

    relations.each do |rel|
      next unless rel.match?(/\w/)
      puts "LICENCE is #{href}\n\n" if rel == "license"
      newlinks << factory.new_link(responsepart: :header, anchor: anchor, href: href, relation: rel, **sections) # parsed['https://example.one.com'][:rel] = "preconnect"
    end
  end
  newlinks
end