/// <param name="dom">Set this parameter to, say, a page that the user just edited, to limit reading to it, so its values don't get overriden by previous pages.
/// Supply the whole dom if nothing has priority (which will mean the data-div will win, because it is first)</param>
/// <param name="collectionSettings"> </param>
/// <param name="updateImgNodeCallback">This is a callback so as not to introduce dependencies on ImageUpdater & the current folder path</param>
public BookData(HtmlDom dom, CollectionSettings collectionSettings, Action<XmlElement> updateImgNodeCallback)
{
_dom = dom;
_updateImgNode = updateImgNodeCallback;
_collectionSettings = collectionSettings;
GetOrCreateDataDiv();
_dataset = GatherDataItemsFromCollectionSettings(_collectionSettings);
GatherDataItemsFromXElement(_dataset,_dom.RawDom);
}
/// <summary>
/// In some cases, we're better off copying from another national language than leaving the field empty.
/// </summary>
/// <remarks>
/// This is a tough decision. Without this, if we have, say, an English Contributors list but English isn't the N1 (L2), then the
/// book won't show it at all. An ideal solution would just order them and then "display the first non-empty one", but that would require some java script... not
/// something could be readily done in CSS, far as I can think.
/// For now, I *think* this won't do any harm, and if it does, it's adding data, not losing it. Users had complained about "losing" the contributor data before.
///</remarks>
private string PossiblyCopyFromAnotherLanguage(XmlElement element, string languageCodeOfTargetField, DataSet data, string key)
{
string classes = element.GetAttribute("class");
if (!string.IsNullOrEmpty(classes))
{
// if this field is normally read-only, make it readable so they can do any translation that might be needed
element.SetAttribute("class", classes.Replace("bloom-readOnlyInTranslationMode", ""));
}
if (!classes.Contains("bloom-copyFromOtherLanguageIfNecessary"))
{
return "";
}
LanguageForm formToCopyFromSinceOursIsMissing = null;
string s = "";
if ((languageCodeOfTargetField == _collectionSettings.Language2Iso639Code || //is it a national language?
languageCodeOfTargetField == _collectionSettings.Language3Iso639Code))
{
formToCopyFromSinceOursIsMissing =
data.TextVariables[key].TextAlternatives.GetBestAlternative(new[] {languageCodeOfTargetField, "*", "en", "fr", "es", "pt"});
if (formToCopyFromSinceOursIsMissing != null)
s = formToCopyFromSinceOursIsMissing.Form;
if (string.IsNullOrEmpty(s))
{
//OK, well even on a non-global language is better than none
//s = data.TextVariables[key].TextAlternatives.GetFirstAlternative();
formToCopyFromSinceOursIsMissing = GetFirstAlternativeForm(data.TextVariables[key].TextAlternatives);
if (formToCopyFromSinceOursIsMissing != null)
s = formToCopyFromSinceOursIsMissing.Form;
}
}
/* this was a fine idea, execpt that if the user then edits it, well, it's not borrowed anymore but we'll still have this sitting there misleading us
//record our dubious deed for posterity
if (formToCopyFromSinceOursIsMissing != null)
{
node.SetAttribute("bloom-languageBloomHadToCopyFrom",
formToCopyFromSinceOursIsMissing.WritingSystemId);
}
*/
return s;
}
/// <summary>
/// Where, for example, somewhere on a page something has data-book='foo' lan='fr',
/// we set the value of that element to French subvalue of the data item 'foo', if we have one.
/// </summary>
private void UpdateDomFromDataSet(DataSet data, string elementName,XmlDocument targetDom)
{
try
{
string query = String.Format("//{0}[(@data-book or @data-collection or @data-library)]", elementName);
XmlNodeList nodesOfInterest = targetDom.SafeSelectNodes(query);
foreach (XmlElement node in nodesOfInterest)
{
string key = node.GetAttribute("data-book").Trim();
if (key == String.Empty)
{
key = node.GetAttribute("data-collection").Trim();
if(key==string.Empty)
{
key = node.GetAttribute("data-library").Trim();
//"library" is the old name for what is now "collection"
}
}
if (!String.IsNullOrEmpty(key) && data.TextVariables.ContainsKey(key))
{
if (node.Name.ToLower() == "img")
{
string imageName =
WebUtility.HtmlDecode(data.TextVariables[key].TextAlternatives.GetFirstAlternative());
string oldImageName = WebUtility.HtmlDecode(node.GetAttribute("src"));
node.SetAttribute("src", imageName);
if (oldImageName != imageName)
{
Guard.AgainstNull(_updateImgNode, "_updateImgNode");
_updateImgNode(node);
}
}
else
{
string lang = node.GetOptionalStringAttribute("lang", "*");
if (lang == "N1" || lang == "N2" || lang == "V")
lang = data.WritingSystemAliases[lang];
// //see comment later about the inability to clear a value. TODO: when we re-write Bloom, make sure this is possible
// if(data.TextVariables[key].TextAlternatives.Forms.Length==0)
// {
// //no text forms == desire to remove it. THe multitextbase prohibits empty strings, so this is the best we can do: completly remove the item.
// targetDom.RemoveChild(node);
// }
// else
if (!String.IsNullOrEmpty(lang)) //if we don't even have this language specified (e.g. no national language), the give up
{
//Ideally, we have this string, in this desired language.
string s = data.TextVariables[key].TextAlternatives.GetBestAlternativeString(new []{lang, "*"});
//But if not, maybe we should copy one in from another national language
if(string.IsNullOrEmpty(s))
s = PossiblyCopyFromAnotherLanguage(node, lang, data, key);
//NB: this was the focus of a multi-hour bug search, and it's not clear that I got it right.
//The problem is that the title page has N1 and n2 alternatives for title, the cover may not.
//the gather page was gathering no values for those alternatives (why not), and so GetBestAlternativeSTring
//was giving "", which we then used to remove our nice values.
//REVIEW: what affect will this have in other pages, other circumstances. Will it make it impossible to clear a value?
//Hoping not, as we are differentiating between "" and just not being in the multitext at all.
//don't overwrite a datadiv alternative with empty just becuase this page has no value for it.
if (s == "" && !data.TextVariables[key].TextAlternatives.ContainsAlternative(lang))
continue;
//hack: until I think of a more elegant way to avoid repeating the language name in N2 when it's the exact same as N1...
if (data.WritingSystemAliases.Count != 0 && lang == data.WritingSystemAliases["N2"] &&
s ==
data.TextVariables[key].TextAlternatives.GetBestAlternativeString(new[]
{
data.
WritingSystemAliases
["N1"]
, "*"
}))
{
s = ""; //don't show it in N2, since it's the same as N1
}
node.InnerXml = s;
//meaning, we'll take "*" if you have it but not the exact choice. * is used for languageName, at least in dec 2011
}
}
}
}
}
catch (Exception error)
{
throw new ApplicationException(
"Error in MakeAllFieldsOfElementTypeConsistent(," + elementName + "). RawDom was:\r\n" +
targetDom.OuterXml, error);
}
}
private static void SendDataToDebugConsole(DataSet data)
{
#if DEBUG
foreach (var item in data.TextVariables)
{
foreach (LanguageForm form in item.Value.TextAlternatives.Forms)
{
Debug.WriteLine("Gathered: {0}[{1}]={2}", item.Key, form.WritingSystemId, form.Form);
}
}
#endif
}
/// <summary>
/// walk throught the sourceDom, collecting up values from elements that have data-book or data-collection attributes.
/// </summary>
private void GatherDataItemsFromXElement(DataSet data, XmlNode sourceElement
/* can be the whole sourceDom or just a page */)
{
string elementName = "*";
try
{
string query = String.Format(".//{0}[(@data-book or @data-library or @data-collection)]", elementName);
XmlNodeList nodesOfInterest = sourceElement.SafeSelectNodes(query);
foreach (XmlElement node in nodesOfInterest)
{
bool isCollectionValue = false;
string key = node.GetAttribute("data-book").Trim();
if (key == String.Empty)
{
key = node.GetAttribute("data-collection").Trim();
if (key == String.Empty)
{
key = node.GetAttribute("data-library").Trim(); //the old (pre-version 1) name of collections was 'library'
}
isCollectionValue = true;
}
string value = node.InnerXml.Trim(); //may contain formatting
if (node.Name.ToLower() == "img")
{
value = node.GetAttribute("src");
//Make the name of the image safe for showing up in raw html (not just in the relatively safe confines of the src attribut),
//becuase it's going to show up between <div> tags. E.g. "Land & Water.png" as the cover page used to kill us.
value = WebUtility.HtmlEncode(WebUtility.HtmlDecode(value));
}
if (!String.IsNullOrEmpty(value) && !value.StartsWith("{"))
//ignore placeholder stuff like "{Book Title}"; that's not a value we want to collect
{
string lang = node.GetOptionalStringAttribute("lang", "*");
if (lang == "") //the above doesn't stop a "" from getting through
lang = "*";
if ((elementName.ToLower() == "textarea" || elementName.ToLower() == "input" ||
node.GetOptionalStringAttribute("contenteditable", "false") == "true") &&
(lang == "V" || lang == "N1" || lang == "N2"))
{
throw new ApplicationException(
"Editable element (e.g. TextArea) should not have placeholder @lang attributes (V,N1,N2)\r\n\r\n" +
node.OuterXml);
}
//if we don't have a value for this variable and this language, add it
if (!data.TextVariables.ContainsKey(key))
{
var t = new MultiTextBase();
t.SetAlternative(lang, value);
data.TextVariables.Add(key, new NamedMutliLingualValue(t, isCollectionValue));
}
else if (!data.TextVariables[key].TextAlternatives.ContainsAlternative(lang))
{
MultiTextBase t = data.TextVariables[key].TextAlternatives;
t.SetAlternative(lang, value);
}
}
}
}
catch (Exception error)
{
throw new ApplicationException(
"Error in GatherDataItemsFromDom(," + elementName + "). RawDom was:\r\n" + sourceElement.OuterXml,
error);
}
}
// records key, lang pairs for which we found an empty element in the source.
/// <summary>
/// walk through the sourceDom, collecting up values from elements that have data-book or data-collection or data-book-attributes attributes.
/// </summary>
private void GatherDataItemsFromXElement(DataSet data,
XmlNode sourceElement, // can be the whole sourceDom or just a page
HashSet<Tuple<string, string>> itemsToDelete = null)
{
string elementName = "*";
try
{
string query = String.Format(".//{0}[(@data-book or @data-library or @data-collection or @data-book-attributes) and not(contains(@class,'bloom-writeOnly'))]", elementName);
XmlNodeList nodesOfInterest = sourceElement.SafeSelectNodes(query);
foreach (XmlElement node in nodesOfInterest)
{
bool isCollectionValue = false;
string key = node.GetAttribute("data-book").Trim();
if (key == String.Empty)
{
key = node.GetAttribute("data-book-attributes").Trim();
if (key != String.Empty)
{
GatherAttributes(data, node, key);
continue;
}
key = node.GetAttribute("data-collection").Trim();
if (key == String.Empty)
{
key = node.GetAttribute("data-library").Trim(); //the old (pre-version 1) name of collections was 'library'
}
isCollectionValue = true;
}
string value;
if (HtmlDom.IsImgOrSomethingWithBackgroundImage(node))
{
value = HtmlDom.GetImageElementUrl(new ElementProxy(node)).UrlEncoded;
KeysOfVariablesThatAreUrlEncoded.Add(key);
}
else
{
var node1 = node.CloneNode(true); // so we can remove labels without modifying node
// Datadiv content should be node content without labels. The labels are not really part
// of the content we want to replicate, they are just information for the user, and
// specific to one context. Also, including them causes them to get repeated in each location;
// SetInnerXmlPreservingLabel() assumes data set content does not include label elements.
var labels = node1.SafeSelectNodes(".//label").Cast<XmlElement>().ToList();
foreach (var label in labels)
label.ParentNode.RemoveChild(label);
value = node1.InnerXml.Trim(); //may contain formatting
if(KeysOfVariablesThatAreUrlEncoded.Contains(key))
{
value = UrlPathString.CreateFromHtmlXmlEncodedString(value).UrlEncoded;
}
}
string lang = node.GetOptionalStringAttribute("lang", "*");
if (lang == "") //the above doesn't stop a "" from getting through
lang = "*";
if (lang == "{V}")
lang = _collectionSettings.Language1Iso639Code;
if(lang == "{N1}")
lang = _collectionSettings.Language2Iso639Code;
if(lang == "{N2}")
lang = _collectionSettings.Language3Iso639Code;
if (string.IsNullOrEmpty(value))
{
// This is a value we may want to delete
if (itemsToDelete != null)
itemsToDelete.Add(Tuple.Create(key, lang));
}
else if (!value.StartsWith("{"))
//ignore placeholder stuff like "{Book Title}"; that's not a value we want to collect
{
if ((elementName.ToLowerInvariant() == "textarea" || elementName.ToLowerInvariant() == "input" ||
node.GetOptionalStringAttribute("contenteditable", "false") == "true") &&
(lang == "V" || lang == "N1" || lang == "N2"))
{
throw new ApplicationException(
"Editable element (e.g. TextArea) should not have placeholder @lang attributes (V,N1,N2)\r\n\r\n" +
node.OuterXml);
}
//if we don't have a value for this variable and this language, add it
if (!data.TextVariables.ContainsKey(key))
{
var t = new MultiTextBase();
t.SetAlternative(lang, value);
data.TextVariables.Add(key, new NamedMutliLingualValue(t, isCollectionValue));
}
else if (!data.TextVariables[key].TextAlternatives.ContainsAlternative(lang))
{
MultiTextBase t = data.TextVariables[key].TextAlternatives;
t.SetAlternative(lang, value);
}
}
if (KeysOfVariablesThatAreUrlEncoded.Contains(key))
{
Debug.Assert(!value.Contains("&"), "In memory, all image urls should be encoded such that & is just &.");
}
}
}
catch (Exception error)
{
throw new ApplicationException(
"Error in GatherDataItemsFromDom(," + elementName + "). RawDom was:\r\n" + sourceElement.OuterXml,
error);
}
}
private static DataSet GatherDataItemsFromCollectionSettings(CollectionSettings collectionSettings)
{
var data = new DataSet();
data.WritingSystemAliases.Add("N1", collectionSettings.Language2Iso639Code);
data.WritingSystemAliases.Add("N2", collectionSettings.Language3Iso639Code);
// if (makeGeneric)
// {
// data.WritingSystemCodes.Add("V", collectionSettings.Language2Iso639Code);
// //This is not an error; we don't want to use the verncular when we're just previewing a book in a non-verncaulr collection
// data.AddGenericLanguageString("iso639Code", collectionSettings.Language1Iso639Code, true);
// //review: maybe this should be, like 'xyz"
// data.AddGenericLanguageString("nameOfLanguage", "(Your Language Name)", true);
// data.AddGenericLanguageString("nameOfNationalLanguage1", "(Region Lang)", true);
// data.AddGenericLanguageString("nameOfNationalLanguage2", "(National Lang)", true);
// data.AddGenericLanguageString("country", "Your Country", true);
// data.AddGenericLanguageString("province", "Your Province", true);
// data.AddGenericLanguageString("district", "Your District", true);
// data.AddGenericLanguageString("languageLocation", "(Language Location)", true);
// }
// else
{
data.WritingSystemAliases.Add("V", collectionSettings.Language1Iso639Code);
data.AddLanguageString("nameOfLanguage", collectionSettings.Language1Name, "*", true);
data.AddLanguageString("nameOfNationalLanguage1",
collectionSettings.GetLanguage2Name(collectionSettings.Language2Iso639Code), "*", true);
data.AddLanguageString("nameOfNationalLanguage2",
collectionSettings.GetLanguage3Name(collectionSettings.Language2Iso639Code), "*", true);
data.UpdateGenericLanguageString("iso639Code", collectionSettings.Language1Iso639Code, true);
data.UpdateGenericLanguageString("country", collectionSettings.Country, true);
data.UpdateGenericLanguageString("province", collectionSettings.Province, true);
data.UpdateGenericLanguageString("district", collectionSettings.District, true);
string location = "";
if (!String.IsNullOrEmpty(collectionSettings.District))
location += collectionSettings.District + @", ";
if (!String.IsNullOrEmpty(collectionSettings.Province))
location += collectionSettings.Province;
location = location.TrimEnd(new[] {' '}).TrimEnd(new[] {','});
if (!String.IsNullOrEmpty(collectionSettings.Country))
{
location += "<br/>" + collectionSettings.Country;
}
data.UpdateGenericLanguageString("languageLocation", location, true);
}
return data;
}
/// <summary>
/// Given a node in the content section of the book that has a data-book attribute, see
/// if this node holds an image and if so, look up the url of the image from the supplied
/// dataset and stick it in there. Handle both img elements and divs that have a
/// background-image in an inline style attribute.
///
/// At the time of this writing, the only image that is handled here is the cover page.
/// The URLs of images in the content of the book are not known to the data-div.
/// But each time the book is loaded up, we collect up data from the xmatter and stick
/// it in the data-div(in the DOM) / dataSet (in an object here in code), stick in a
/// blank xmatter, then push the values back into the xmatter.
/// </summary>
/// <returns>true if this node is an image holder of some sort.</returns>
private bool UpdateImageFromDataSet(DataSet data, XmlElement node, string key)
{
if (!HtmlDom.IsImgOrSomethingWithBackgroundImage(node))
return false;
var newImageUrl = UrlPathString.CreateFromUrlEncodedString(data.TextVariables[key].TextAlternatives.GetFirstAlternative());
var oldImageUrl = HtmlDom.GetImageElementUrl(node);
var imgOrDivWithBackgroundImage = new ElementProxy(node);
HtmlDom.SetImageElementUrl(imgOrDivWithBackgroundImage,newImageUrl);
if (!newImageUrl.Equals(oldImageUrl))
{
Guard.AgainstNull(_updateImgNode, "_updateImgNode");
try
{
_updateImgNode(node);
}
catch (TagLib.CorruptFileException e)
{
NonFatalProblem.Report(ModalIf.Beta, PassiveIf.All, "Problem reading image metadata", newImageUrl.NotEncoded, e);
return false;
}
}
return true;
}
public Metadata GetLicenseMetadata()
{
var data = new DataSet();
GatherDataItemsFromXElement(data, _dom.RawDom);
var metadata = new Metadata();
NamedMutliLingualValue d;
if (data.TextVariables.TryGetValue("copyright", out d))
{
metadata.CopyrightNotice = WebUtility.HtmlDecode(d.TextAlternatives.GetFirstAlternative());
}
string licenseUrl = "";
if (data.TextVariables.TryGetValue("licenseUrl", out d))
{
licenseUrl = d.TextAlternatives.GetFirstAlternative();
}
//Enhance: have a place for notes (amendments to license). It's already in the frontmatter, under "licenseNotes"
if (licenseUrl == null || licenseUrl.Trim() == "")
{
//NB: we are mapping "RightsStatement" (which comes from XMP-dc:Rights) to "LicenseNotes" in the html.
//custom licenses live in this field
if (data.TextVariables.TryGetValue("licenseNotes", out d))
{
string licenseNotes = d.TextAlternatives.GetFirstAlternative();
metadata.License = new CustomLicense {RightsStatement = licenseNotes};
}
else
{
//how to detect a null license was chosen? We're using the fact that it has a description, but nothing else.
if (data.TextVariables.TryGetValue("licenseDescription", out d))
{
metadata.License = new NullLicense(); //"contact the copyright owner
}
else
{
//looks like the first time. Nudge them with a nice default
metadata.License = new CreativeCommonsLicense(true, true,
CreativeCommonsLicense.DerivativeRules.Derivatives);
}
}
}
else
{
metadata.License = CreativeCommonsLicense.FromLicenseUrl(licenseUrl);
}
return metadata;
}
private void UpdateAttributes(DataSet data, XmlElement node, string key)
{
List<KeyValuePair<string, string>> attributes;
if (!data.Attributes.TryGetValue(key, out attributes))
return;
foreach (var attribute in attributes)
node.SetAttribute(attribute.Key, attribute.Value);
}
/// <summary>
/// Where, for example, somewhere on a page something has data-book='foo' lang='fr',
/// we set the value of that element to French subvalue of the data item 'foo', if we have one.
/// </summary>
private void UpdateDomFromDataSet(DataSet data, string elementName,XmlDocument targetDom, HashSet<Tuple<string, string>> itemsToDelete)
{
try
{
var query = String.Format("//{0}[(@data-book or @data-collection or @data-library or @data-book-attributes)]", elementName);
var nodesOfInterest = targetDom.SafeSelectNodes(query);
foreach (XmlElement node in nodesOfInterest)
{
var key = node.GetAttribute("data-book").Trim();
if (key == string.Empty)
{
key = node.GetAttribute("data-book-attributes").Trim();
if (key != string.Empty)
{
UpdateAttributes(data, node, key);
continue;
}
key = node.GetAttribute("data-collection").Trim();
if (key == string.Empty)
{
key = node.GetAttribute("data-library").Trim(); //"library" is the old name for what is now "collection"
}
}
if (string.IsNullOrEmpty(key)) continue;
if (data.TextVariables.ContainsKey(key))
{
if (UpdateImageFromDataSet(data, node, key)) continue;
var lang = node.GetOptionalStringAttribute("lang", "*");
if (lang == "N1" || lang == "N2" || lang == "V")
lang = data.WritingSystemAliases[lang];
// //see comment later about the inability to clear a value. TODO: when we re-write Bloom, make sure this is possible
// if(data.TextVariables[key].TextAlternatives.Forms.Length==0)
// {
// //no text forms == desire to remove it. THe multitextbase prohibits empty strings, so this is the best we can do: completly remove the item.
// targetDom.RemoveChild(node);
// }
// else
if (!string.IsNullOrEmpty(lang)) //if we don't even have this language specified (e.g. no national language), the give up
{
//Ideally, we have this string, in this desired language.
var s = data.TextVariables[key].TextAlternatives.GetBestAlternativeString(new[] {lang, "*"});
if(KeysOfVariablesThatAreUrlEncoded.Contains(key))
{
Debug.Assert(!s.Contains("&"),"In memory, all image urls should be encoded such that & is just &.");
}
//But if not, maybe we should copy one in from another national language
if (string.IsNullOrEmpty(s))
s = PossiblyCopyFromAnotherLanguage(node, lang, data, key);
//NB: this was the focus of a multi-hour bug search, and it's not clear that I got it right.
//The problem is that the title page has N1 and n2 alternatives for title, the cover may not.
//the gather page was gathering no values for those alternatives (why not), and so GetBestAlternativeSTring
//was giving "", which we then used to remove our nice values.
//REVIEW: what affect will this have in other pages, other circumstances. Will it make it impossible to clear a value?
//Hoping not, as we are differentiating between "" and just not being in the multitext at all.
//don't overwrite a datadiv alternative with empty just becuase this page has no value for it.
if (s == "" && !data.TextVariables[key].TextAlternatives.ContainsAlternative(lang))
continue;
//hack: until I think of a more elegant way to avoid repeating the language name in N2 when it's the exact same as N1...
if (data.WritingSystemAliases.Count != 0 && lang == data.WritingSystemAliases["N2"] &&
s ==
data.TextVariables[key].TextAlternatives.GetBestAlternativeString(new[]
{
data.
WritingSystemAliases
["N1"]
, "*"
}))
{
s = ""; //don't show it in N2, since it's the same as N1
}
SetInnerXmlPreservingLabel(key, node, s);
}
}
else if (!HtmlDom.IsImgOrSomethingWithBackgroundImage(node))
{
// See whether we need to delete something
var lang = node.GetOptionalStringAttribute("lang", "*");
if (lang == "N1" || lang == "N2" || lang == "V")
lang = data.WritingSystemAliases[lang];
if (itemsToDelete.Contains(Tuple.Create(key, lang)))
{
SetInnerXmlPreservingLabel(key, node, "");// a later process may remove node altogether.
}
}
}
}
catch (Exception error)
{
throw new ApplicationException(
"Error in UpdateDomFromDataSet(," + elementName + "). RawDom was:\r\n" +
targetDom.OuterXml, error);
}
}
/// <summary>
/// Topics are uni-directional value, react™-style. The UI tells the book to change the topic
/// key, and then eventually the page/book is re-evaluated and the appropriate topic is displayed
/// on the page.
/// To differentiate from fields with @data-book, which are two-way, the topic on the page instead
/// has a @data-derived attribute (in the data-div, it is still a data-book... perhaps that too could
/// change to something like data-book-source, but it's not clear to me yet, so.. not yet).
/// When the topic is changed, the javascript sends c# a message with the new English Key for the topic is set in the data-div,
/// and then the page is re-computed. That leads to this method, which grabs the
/// english topic (which serves as the 'key') from the datadiv. It then finds the placeholder
/// for the topic and fills it with the best translation it can find.
/// </summary>
private void SetUpDisplayOfTopicInBook(DataSet data)
{
var topicPageElement = this._dom.SelectSingleNode("//div[@data-derived='topic']");
if (topicPageElement == null)
{
//old-style. here we don't have the data-derived, so we need to avoid picking from the datadiv
topicPageElement = this._dom.SelectSingleNode("//div[not(id='bloomDataDiv')]//div[@data-book='topic']");
if (topicPageElement == null)
{
//most unit tests do not have complete books, so this not surprising. It just means we don't have anything to do
return;
}
}
//clear it out what's there now
topicPageElement.RemoveAttribute("lang");
topicPageElement.InnerText = "";
NamedMutliLingualValue topicData;
var parentOfTopicDisplayElement = ((XmlElement)(topicPageElement.ParentNode));
//this just lets us have css rules that vary if there is a topic (allows other text to be centered instead left-aligned)
//we'll change it later if we find there is a topic
parentOfTopicDisplayElement.SetAttribute("data-have-topic", "false");
//if we have no topic element in the data-div
//leave the field in the page with an empty text.
if (!data.TextVariables.TryGetValue("topic", out topicData))
{
return;
}
//we use English as the "key" for topics.
var englishTopic = topicData.TextAlternatives.GetExactAlternative("en");
//if we have no topic, just clear it out from the page
if (string.IsNullOrEmpty(englishTopic) || englishTopic == "NoTopic")
return;
parentOfTopicDisplayElement.SetAttribute("data-have-topic", "true");
var stringId = "Topics." + englishTopic;
//get the topic in the most prominent language for which we have a translation
var langOfTopicToShowOnCover = _collectionSettings.Language1Iso639Code;
if (LocalizationManager.GetIsStringAvailableForLangId(stringId, _collectionSettings.Language1Iso639Code))
{
langOfTopicToShowOnCover = _collectionSettings.Language1Iso639Code;
}
else if (LocalizationManager.GetIsStringAvailableForLangId(stringId, _collectionSettings.Language2Iso639Code))
{
langOfTopicToShowOnCover = _collectionSettings.Language2Iso639Code;
}
else if (LocalizationManager.GetIsStringAvailableForLangId(stringId, _collectionSettings.Language3Iso639Code))
{
langOfTopicToShowOnCover = _collectionSettings.Language3Iso639Code;
}
else
{
langOfTopicToShowOnCover = "en";
}
var bestTranslation = LocalizationManager.GetDynamicStringOrEnglish("Bloom", stringId, englishTopic,
"this is a book topic", langOfTopicToShowOnCover);
//NB: in a unit test environment, GetDynamicStringOrEnglish is going to give us the id back, which is annoying.
if (bestTranslation == stringId)
bestTranslation = englishTopic;
topicPageElement.SetAttribute("lang", langOfTopicToShowOnCover);
topicPageElement.InnerText = bestTranslation;
}
/// <summary>
/// In some cases, we're better off copying from another national language than leaving the field empty.
/// </summary>
/// <remarks>
/// This is a tough decision. Without this, if we have, say, an English Contributors list but English isn't the N1 (L2), then the
/// book won't show it at all. An ideal solution would just order them and then "display the first non-empty one", but that would require some java script... not
/// something could be readily done in CSS, far as I can think.
/// For now, I *think* this won't do any harm, and if it does, it's adding data, not losing it. Users had complained about "losing" the contributor data before.
///</remarks>
private string PossiblyCopyFromAnotherLanguage(XmlElement element, string languageCodeOfTargetField, DataSet data, string key)
{
string classes = element.GetAttribute("class");
if (!string.IsNullOrEmpty(classes))
{
// if this field is normally read-only, make it readable so they can do any translation that might be needed
element.SetAttribute("class", classes.Replace("bloom-readOnlyInTranslationMode", ""));
}
if (!classes.Contains("bloom-copyFromOtherLanguageIfNecessary"))
{
return "";
}
LanguageForm formToCopyFromSinceOursIsMissing = null;
string s = "";
if ((languageCodeOfTargetField == _collectionSettings.Language2Iso639Code || //is it a national language?
languageCodeOfTargetField == _collectionSettings.Language3Iso639Code) ||
//this one is a kludge as we clearly have this case of a vernacular field that people have used
//to hold stuff that should be copied to every shell. So we can either remove the restriction of the
//first two clauses in this if statement, or add another bloom-___ class in order to make execptions.
//Today, I'm not seing the issue clearly enough, so I'm just going to path this one exisint hole.
classes.Contains("smallCoverCredits"))
{
formToCopyFromSinceOursIsMissing =
data.TextVariables[key].TextAlternatives.GetBestAlternative(new[] {languageCodeOfTargetField, "*", "en", "fr", "es", "pt"});
if (formToCopyFromSinceOursIsMissing != null)
s = formToCopyFromSinceOursIsMissing.Form;
if (string.IsNullOrEmpty(s))
{
//OK, well even on a non-global language is better than none
//s = data.TextVariables[key].TextAlternatives.GetFirstAlternative();
formToCopyFromSinceOursIsMissing = GetFirstAlternativeForm(data.TextVariables[key].TextAlternatives);
if (formToCopyFromSinceOursIsMissing != null)
s = formToCopyFromSinceOursIsMissing.Form;
}
}
/* this was a fine idea, execpt that if the user then edits it, well, it's not borrowed anymore but we'll still have this sitting there misleading us
//record our dubious deed for posterity
if (formToCopyFromSinceOursIsMissing != null)
{
node.SetAttribute("bloom-languageBloomHadToCopyFrom",
formToCopyFromSinceOursIsMissing.WritingSystemId);
}
*/
return s;
}
private void MigrateData(DataSet data)
{
//Until late in Bloom 3, we collected the topic in the National language, which is messy because then we would have to know how to
//translate from all those languages to all other languages. Now, we just save English, and translate from English to whatever.
//By far the largest number of books posted to bloomlibrary with this problem were Tok Pisin books, which actually just had
//an English word as their value for "topic", so there we just switch it over to English.
NamedMutliLingualValue topic;
if (!data.TextVariables.TryGetValue("topic", out topic))
return;
var topicStrings = topic.TextAlternatives;
if (string.IsNullOrEmpty(topicStrings["en"] ) && topicStrings["tpi"] != null)
{
topicStrings["en"] = topicStrings["tpi"];
topicStrings.RemoveLanguageForm(topicStrings.Find("tpi"));
}
// BL-2746 For awhile during the v3.3 beta period, after the addition of ckeditor
// our topic string was getting wrapped in html paragraph markers. There were a good
// number of beta testers, so we need to clean up that mess.
topicStrings.Forms
.ForEach(
languageForm =>
topicStrings[languageForm.WritingSystemId] = languageForm.Form.Replace("<p>", "").Replace("</p>", ""));
if (!string.IsNullOrEmpty(topicStrings["en"]))
{
//starting with 3.5, we only store the English key in the datadiv.
topicStrings.Forms
.Where(lf => lf.WritingSystemId != "en")
.ForEach(lf => topicStrings.RemoveLanguageForm(lf));
_dom.SafeSelectNodes("//div[@id='bloomDataDiv']/div[@data-book='topic' and not(@lang='en')]")
.Cast<XmlElement>()
.ForEach(e => e.ParentNode.RemoveChild(e));
}
}
private void InjectXMatter(string initialPath, BookStorage storage, Layout sizeAndOrientation)
{
//now add in the xmatter from the currently selected xmatter pack
if (!TestingSoSkipAddingXMatter)
{
var data = new DataSet();
Debug.Assert(!string.IsNullOrEmpty(_collectionSettings.Language1Iso639Code));
Debug.Assert(!string.IsNullOrEmpty(_collectionSettings.Language2Iso639Code));
data.WritingSystemCodes.Add("V", _collectionSettings.Language1Iso639Code);
data.WritingSystemCodes.Add("N1", _collectionSettings.Language2Iso639Code);
data.WritingSystemCodes.Add("N2", _collectionSettings.Language3Iso639Code);
//by default, this comes from the collection, but the book can select one, inlucing "null" to select the factory-supplied empty xmatter
var xmatterName = storage.Dom.GetMetaValue("xmatter", _collectionSettings.XMatterPackName);
var helper = new XMatterHelper(storage.Dom, xmatterName, _fileLocator);
helper.FolderPathForCopyingXMatterFiles = storage.FolderPath;
helper.InjectXMatter(data.WritingSystemCodes, sizeAndOrientation);
}
}
public void SetLicenseMetdata(Metadata metadata)
{
var data = new DataSet();
GatherDataItemsFromXElement(data, _dom.RawDom);
string copyright = metadata.CopyrightNotice;
data.UpdateLanguageString("copyright", copyright, "*", false);
string description = metadata.License.GetDescription("en");
data.UpdateLanguageString("licenseDescription", description, "en", false);
string licenseUrl = metadata.License.Url;
data.UpdateLanguageString("licenseUrl", licenseUrl, "*", false);
string licenseNotes = metadata.License.RightsStatement;
data.UpdateLanguageString("licenseNotes", licenseNotes, "*", false);
string licenseImageName = metadata.License.GetImage() == null ? "" : "license.png";
data.UpdateGenericLanguageString("licenseImage", licenseImageName, false);
UpdateDomFromDataSet(data, "*", _dom.RawDom);
//UpdateDomFromDataSet() is not able to remove items yet, so we do it explicity
RemoveDataDivElementIfEmptyValue("licenseDescription", description);
RemoveDataDivElementIfEmptyValue("licenseImage", licenseImageName);
RemoveDataDivElementIfEmptyValue("licenseUrl", licenseUrl);
RemoveDataDivElementIfEmptyValue("copyright", copyright);
RemoveDataDivElementIfEmptyValue("licenseNotes", licenseNotes);
}
public void Setup()
{
_dom = new HtmlDom(@"<html><head> <link href='file://blahblah\\a5portrait.css' type='text/css' /></head><body><div id='bloomDataDiv'></div><div id ='firstPage' class='bloom-page'>1st page</div></body></html>");
_dataSet = new DataSet();
_dataSet.WritingSystemCodes.Add("V","xyz");
_dataSet.WritingSystemCodes.Add("N1", "fr");
_dataSet.WritingSystemCodes.Add("N2", "en");
}
private void GatherAttributes(DataSet data, XmlElement node, string key)
{
if (data.Attributes.ContainsKey(key))
return;
List<KeyValuePair<string, string>> attributes = new List<KeyValuePair<string, string>>();
foreach (XmlAttribute attribute in node.Attributes)
{
if (attribute.Name != "data-book-attributes")
attributes.Add(new KeyValuePair<string, string>(attribute.Name, attribute.Value));
}
data.Attributes.Add(key, attributes);
}