Fuzzy anchoring on unrendered PDF pages

Also change the PDF anchoring to leverage the HTML anchoring by
allowing the options to be passed into the module-level anchor
functions in the pdf and html modules so that the pdf module can
be only responsible for finding the appropriate page and then
delegating.

h/static/scripts/annotator/anchoring/html.coffee

@@ -16,11 +16,7 @@
 # :return: A Promise that resolves to a Range on success.
 # :rtype: Promise
 ####
-exports.anchor = (selectors) ->
-  options =
-    root: document.body
-    ignoreSelector: '[class^="annotator-"]'
-
+exports.anchor = (selectors, options = {}) ->
   # Selectors
   fragment = null
   position = null

h/static/scripts/annotator/anchoring/pdf.coffee

@@ -3,10 +3,8 @@ seek = require('dom-seek')
 Annotator = require('annotator')
 xpathRange = Annotator.Range
 
-{
-  TextPositionAnchor
-  TextQuoteAnchor
-} = require('./types')
+html = require('./html')
+{TextPositionAnchor, TextQuoteAnchor} = require('./types')
 
 
 getSiblingIndex = (node) ->
@@ -77,11 +75,7 @@ findPage = (offset) ->
 # :return: A Promise that resolves to a Range on success.
 # :rtype: Promise
 ####
-exports.anchor = (selectors) ->
-  options =
-    root: document.getElementById('viewer')
-    ignoreSelector: '[class^="annotator-"]'
-
+exports.anchor = (selectors, options = {}) ->
   # Selectors
   position = null
   quote = null
@@ -104,24 +98,13 @@ exports.anchor = (selectors) ->
     else
       return range
 
-  if position?
-    promise = promise.catch ->
-      findPage(position.start)
-      .then(({index, offset, textContent}) ->
-        page = getPage(index)
-
-        start = position.start - offset
-        end = position.end - offset
-        length = end - start
-
-        assertQuote(textContent.substr(start, length))
-
+  anchorByPosition = (page, anchor) ->
     renderingState = page.renderingState
     renderingDone = page.textLayer?.renderingDone
     if renderingState is RenderingStates.FINISHED and renderingDone
       root = page.textLayer.textLayerDiv
-          Promise.resolve(TextPositionAnchor.fromSelector({start, end}, {root}))
-          .then((a) -> Promise.resolve(a.toRange({root})))
+      selector = anchor.toSelector()
+      return html.anchor([selector], {root})
     else
       div = page.div ? page.el
       placeholder = div.getElementsByClassName('annotator-placeholder')[0]
@@ -134,7 +117,18 @@ exports.anchor = (selectors) ->
       range.setStartBefore(placeholder)
       range.setEndAfter(placeholder)
       return range
-      )
+
+  if position?
+    promise = promise.catch ->
+      return findPage(position.start)
+      .then ({index, offset, textContent}) ->
+        page = getPage(index)
+        start = position.start - offset
+        end = position.end - offset
+        length = end - start
+        assertQuote(textContent.substr(start, length))
+        anchor = new TextPositionAnchor(start, end)
+        return anchorByPosition(page, anchor)
 
   if quote?
     promise = promise.catch ->
@@ -142,30 +136,26 @@ exports.anchor = (selectors) ->
 
       pageSearches = for pageIndex in [0...pagesCount]
         page = getPage(pageIndex)
-        continue unless page.textLayer?.renderingDone
-
         content = getPageTextContent(pageIndex)
         offset = getPageOffset(pageIndex)
-
-        Promise.all([content, offset, page]).then((results) ->
+        Promise.all([content, offset, page]).then (results) ->
           [content, offset, page] = results
-          quoteOptions = {root: page.textLayer.textLayerDiv}
+          pageOptions = {root: {textContent: content}}
           if position?
             # XXX: must be on one page
             start = position.start - offset
             end = position.end - offset
-            quoteOptions.position = {start, end}
-
-          return TextQuoteAnchor
-          .fromSelector(quote, quoteOptions)
-          .toRange(quoteOptions)
-        ).catch(-> null)
-
-      return Promise.all(pageSearches).then((results) ->
+            pageOptions.position = {start, end}
+          anchor = new TextQuoteAnchor.fromSelector(quote, pageOptions)
+          return Promise.resolve(anchor)
+          .then((a) -> return a.toPositionAnchor(pageOptions))
+          .then((a) -> return anchorByPosition(page, a))
+
+      pageSearches = (p.catch(-> null) for p in pageSearches)
+      return Promise.all(pageSearches).then (results) ->
         for result in results when result?
           return result
         throw new Error('quote not found')
-      )
 
   return promise
 

h/static/scripts/annotator/anchoring/types.coffee

@@ -211,6 +211,18 @@ class TextQuoteAnchor extends Anchor
     return new TextQuoteAnchor(exact, prefix, suffix, start, end)
 
   toRange: (options = {}) ->
+    return this.toPositionAnchor(options).toRange()
+
+  toSelector: ->
+    selector = {
+      type: 'TextQuoteSelector'
+      exact: @quote
+    }
+    if @prefix? then selector.prefix = @prefix
+    if @suffix? then selector.suffix = @suffix
+    return selector
+
+  toPositionAnchor: (options = {}) ->
     root = options.root or document.body
     dmp = new DiffMatchPatch()
 
@@ -245,16 +257,7 @@ class TextQuoteAnchor extends Anchor
     dmp.Match_Distance = 64
     {start, end} = slices.reduce(foldSlices, {start, end, loc})
 
-    return new TextPositionAnchor(start, end).toRange(options)
-
-  toSelector: ->
-    selector = {
-      type: 'TextQuoteSelector'
-      exact: @quote
-    }
-    if @prefix? then selector.prefix = @prefix
-    if @suffix? then selector.suffix = @suffix
-    return selector
+    return new TextPositionAnchor(start, end)
 
 exports.Anchor = Anchor
 exports.FragmentAnchor = FragmentAnchor

h/static/scripts/annotator/guest.coffee

@@ -155,8 +155,9 @@ module.exports = class Guest extends Annotator
       deadHighlights = []
 
       _anchor = (target) ->
+        options = {ignoreSelector: '[class^="annotator-"]'}
         return new Promise(raf)
-        .then(-> anchoring.anchor(target.selector))
+        .then(-> anchoring.anchor(target.selector, options))
         .then((range) -> {annotation, target, range})
         .catch(-> {annotation, target})