Theodor-Springmann-Stiftung/hamann-ausgabe-core
1
0
Fork 0
mirror of https://github.com/Theodor-Springmann-Stiftung/hamann-ausgabe-core.git synced 2025-10-30 01:35:32 +00:00
Code Issues Packages Projects Releases Wiki Activity

Added generic collections of objects. TODO: deserialization of metadata

Browse Source
This commit is contained in:
schnulller
2022-06-19 23:20:35 +02:00
parent d8a4fe7790
commit ba3d63ffba
17 changed files with 175 additions and 75 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
HaWeb/XMLParser/IXMLRoot.cs
View File

@@ -13,7 +13,15 @@ public interface IXMLRoot {
// XPaths to determine if container is present
public abstract string[] XPathContainer { get; }
// Tag Name of child objects to be collected
// Collections of Elements to be created from this Root
// Key: the key under which the element(s) will be files
// xPath: the (absolute) XPath to the element(s)
// KeyFunc: How to extrect an identifier for the single element in the collection
// Searchable: Will the element be indexed for full-text-search?
public abstract (string Key, string xPath, Func<XElement, string?> KeyFunc, bool Searchable)[]? XPathCollection { get; }
// Determines child objects to be collected
// (deprecated see collections above; only used internally)
public abstract Predicate<XElement> IsCollectedObject { get; }
// Gets the Key of a collected object

5
HaWeb/XMLParser/IXMLService.cs
View File

@@ -2,12 +2,13 @@ namespace HaWeb.XMLParser;
using System.Xml.Linq;
using Microsoft.AspNetCore.Mvc.ModelBinding;
using HaWeb.Models;
using HaXMLReader.Interfaces;
public interface IXMLService {
public IXMLRoot? GetRoot(string name);
public List<IXMLRoot>? GetRootsList();
public Dictionary<string, IXMLRoot>? GetRootsDictionary();
public List<XMLRootDocument>? ProbeHamannFile(XDocument document, ModelStateDictionary ModelState);
public List<XMLRootDocument>? ProbeFile(XDocument document, ModelStateDictionary ModelState);
public Dictionary<string, FileList?>? GetUsedDictionary();
public XElement? MergeUsedDocuments(ModelStateDictionary ModelState);
public void Use(XMLRootDocument doc);
@@ -17,4 +18,6 @@ public interface IXMLService {
public void UnUse(string prefix);
public void UnUseProduction();
public void SetInProduction();
public void SetInProduction(XDocument document);
public List<(string Index, List<(string Page, string Line, string Preview)> Results)>? SearchCollection(string collection, string searchword, IReaderService reader);
}

65
HaWeb/XMLParser/XMLService.cs
View File

@@ -1,7 +1,13 @@
namespace HaWeb.XMLParser;
using System.Xml.Linq;
using System.Xml.XPath;
using Microsoft.AspNetCore.Mvc.ModelBinding;
using HaWeb.Models;
using HaWeb.SearchHelpers;
using System.Collections.Concurrent;
using System.Threading.Tasks;
using System.Text;
using HaXMLReader.Interfaces;
public class XMLService : IXMLService {
private Dictionary<string, FileList?>? _Used;
@@ -9,6 +15,9 @@ public class XMLService : IXMLService {
private Stack<Dictionary<string, FileList?>>? _InProduction;
private Dictionary<string, Dictionary<string, CollectedItem>> _collectedProduction;
private Dictionary<string, Dictionary<string, CollectedItem>> _collectedUsed;
public XMLService() {
// Getting all classes which implement IXMLRoot for possible document endpoints
var types = _GetAllTypesThatImplementInterface<IXMLRoot>().ToList();
@@ -50,9 +59,63 @@ public class XMLService : IXMLService {
_InProduction.Push(inProduction);
}
public void SetInProduction(XDocument document) {
if (document == null || _Roots == null) return;
var ret = new ConcurrentDictionary<string, ConcurrentDictionary<string, CollectedItem>>();
Parallel.ForEach(_Roots, (root) => {
if (root.Value.XPathCollection != null)
foreach (var coll in root.Value.XPathCollection) {
var elem = document.XPathSelectElements(coll.xPath);
if (elem != null && elem.Any()) {
if (!ret.ContainsKey(coll.Key))
ret[coll.Key] = new ConcurrentDictionary<string, CollectedItem>();
foreach(var e in elem) {
var k = coll.KeyFunc(e);
if (k != null) {
var searchtext = coll.Searchable ?
StringHelpers.NormalizeWhiteSpace(e.ToString(), ' ', false) :
null;
ret[coll.Key][k] = new CollectedItem(k, e, root.Value, coll.Key, searchtext);
}
}
}
}
});
_collectedProduction = ret.ToDictionary(x => x.Key, y => y.Value.ToDictionary(z => z.Key, f => f.Value, null), null);
}
public List<(string Index, List<(string Page, string Line, string Preview)> Results)>? SearchCollection(string collection, string searchword, IReaderService reader) {
if (!_collectedProduction.ContainsKey(collection)) return null;
var searchableObjects = _collectedProduction[collection];
var res = new ConcurrentBag<(string Index, List<(string Page, string Line, string preview)> Results)>();
var sw = StringHelpers.NormalizeWhiteSpace(searchword.Trim());
Parallel.ForEach(searchableObjects, (obj) => {
if (obj.Value.SearchText != null) {
var state = new SearchState(sw);
var rd = reader.RequestStringReader(obj.Value.SearchText);
var parser = new HaWeb.HTMLParser.LineXMLHelper<SearchState>(state, rd, new StringBuilder(), null, null, null, SearchRules.TextRules, SearchRules.WhitespaceRules);
rd.Read();
if (state.Results != null)
res.Add((
obj.Value.Index,
state.Results.Select(x => (
x.Page,
x.Line,
parser.Lines != null ?
parser.Lines
.Where(y => y.Page == x.Page && y.Line == x.Line)
.Select(x => x.Text)
.FirstOrDefault(string.Empty)
: ""
)).ToList()));
}
});
return res.ToList();
}
public void UnUseProduction() => this._InProduction = null;
public List<XMLRootDocument>? ProbeHamannFile(XDocument document, ModelStateDictionary ModelState) {
public List<XMLRootDocument>? ProbeFile(XDocument document, ModelStateDictionary ModelState) {
if (document.Root!.Name != "opus") {
ModelState.AddModelError("Error", "A valid Hamann-Docuemnt must begin with <opus>");
return null;