some css tweaks regarding colors

[wolnelektury.git] / apps / search / index.py

diff --git a/apps/search/index.py b/apps/search/index.py
index 25c16a9..6c76901 100644 (file)
--- a/apps/search/index.py
+++ b/apps/search/index.py
@@ -6,10 +6,10 @@ from lucene import SimpleFSDirectory, IndexWriter, CheckIndex, \
     NumericField, Version, Document, JavaError, IndexSearcher, \
     QueryParser, PerFieldAnalyzerWrapper, \
     SimpleAnalyzer, PolishAnalyzer, ArrayList, \
     NumericField, Version, Document, JavaError, IndexSearcher, \
     QueryParser, PerFieldAnalyzerWrapper, \
     SimpleAnalyzer, PolishAnalyzer, ArrayList, \

-    KeywordAnalyzer, NumericRangeQuery, BooleanQuery, \
-    BlockJoinQuery, BlockJoinCollector, TermsFilter, \
+    KeywordAnalyzer, NumericRangeQuery, NumericRangeFilter, BooleanQuery, \
+    BlockJoinQuery, BlockJoinCollector, Filter, TermsFilter, ChainedFilter, \

     HashSet, BooleanClause, Term, CharTermAttribute, \
     HashSet, BooleanClause, Term, CharTermAttribute, \

-    PhraseQuery, MultiPhraseQuery, StringReader, TermQuery, BlockJoinQuery, \
+    PhraseQuery, MultiPhraseQuery, StringReader, TermQuery, \

     FuzzyQuery, FuzzyTermEnum, PrefixTermEnum, Sort, Integer, \
     SimpleHTMLFormatter, Highlighter, QueryScorer, TokenSources, TextFragment, \
     BooleanFilter, TermsFilter, FilterClause, QueryWrapperFilter, \
     FuzzyQuery, FuzzyTermEnum, PrefixTermEnum, Sort, Integer, \
     SimpleHTMLFormatter, Highlighter, QueryScorer, TokenSources, TextFragment, \
     BooleanFilter, TermsFilter, FilterClause, QueryWrapperFilter, \


@@ -25,7 +25,7 @@ import re
 import errno
 from librarian import dcparser
 from librarian.parser import WLDocument
 import errno
 from librarian import dcparser
 from librarian.parser import WLDocument

-from catalogue import models
+import catalogue.models

 from multiprocessing.pool import ThreadPool
 from threading import current_thread
 import atexit
 from multiprocessing.pool import ThreadPool
 from threading import current_thread
 import atexit


@@ -35,11 +35,9 @@ import traceback
 class WLAnalyzer(PerFieldAnalyzerWrapper):
     def __init__(self):
         polish = PolishAnalyzer(Version.LUCENE_34)
 class WLAnalyzer(PerFieldAnalyzerWrapper):
     def __init__(self):
         polish = PolishAnalyzer(Version.LUCENE_34)

-        polish_gap = PolishAnalyzer(Version.LUCENE_34)

         #        polish_gap.setPositionIncrementGap(999)
 
         simple = SimpleAnalyzer(Version.LUCENE_34)
         #        polish_gap.setPositionIncrementGap(999)
 
         simple = SimpleAnalyzer(Version.LUCENE_34)

-        simple_gap = SimpleAnalyzer(Version.LUCENE_34)

         #        simple_gap.setPositionIncrementGap(999)
 
         keyword = KeywordAnalyzer(Version.LUCENE_34)
         #        simple_gap.setPositionIncrementGap(999)
 
         keyword = KeywordAnalyzer(Version.LUCENE_34)


@@ -48,21 +46,24 @@ class WLAnalyzer(PerFieldAnalyzerWrapper):
 
         PerFieldAnalyzerWrapper.__init__(self, polish)
 
 
         PerFieldAnalyzerWrapper.__init__(self, polish)
 

-        self.addAnalyzer("tags", simple_gap)
+        self.addAnalyzer("tags", simple)

         self.addAnalyzer("technical_editors", simple)
         self.addAnalyzer("editors", simple)
         self.addAnalyzer("url", keyword)
         self.addAnalyzer("source_url", keyword)
         self.addAnalyzer("source_name", simple)
         self.addAnalyzer("publisher", simple)
         self.addAnalyzer("technical_editors", simple)
         self.addAnalyzer("editors", simple)
         self.addAnalyzer("url", keyword)
         self.addAnalyzer("source_url", keyword)
         self.addAnalyzer("source_name", simple)
         self.addAnalyzer("publisher", simple)

-        self.addAnalyzer("author", simple)
+        self.addAnalyzer("authors", simple)

         self.addAnalyzer("is_book", keyword)
         self.addAnalyzer("is_book", keyword)

+        # shouldn't the title have two forms? _pl and simple?

 
 

-        self.addAnalyzer("themes", simple_gap)
-        self.addAnalyzer("themes_pl", polish_gap)
+        self.addAnalyzer("themes", simple)
+        self.addAnalyzer("themes_pl", polish)

 
 

-        self.addAnalyzer("tag_name", simple_gap)
-        self.addAnalyzer("tag_name_pl", polish_gap)
+        self.addAnalyzer("tag_name", simple)
+        self.addAnalyzer("tag_name_pl", polish)
+
+        self.addAnalyzer("translators", simple)

 
         self.addAnalyzer("KEYWORD", keyword)
         self.addAnalyzer("SIMPLE", simple)
 
         self.addAnalyzer("KEYWORD", keyword)
         self.addAnalyzer("SIMPLE", simple)


@@ -114,15 +115,17 @@ class Snippets(object):
         return self
 
     def add(self, snippet):
         return self
 
     def add(self, snippet):

-        l = len(snippet)
-        self.file.write(snippet.encode('utf-8'))
+        txt = snippet.encode('utf-8')
+        l = len(txt)
+        self.file.write(txt)

         pos = (self.position, l)
         self.position += l
         return pos
 
     def get(self, pos):
         self.file.seek(pos[0], 0)
         pos = (self.position, l)
         self.position += l
         return pos
 
     def get(self, pos):
         self.file.seek(pos[0], 0)

-        return self.read(pos[1]).decode('utf-8')
+        txt = self.file.read(pos[1]).decode('utf-8')
+        return txt

 
     def close(self):
         self.file.close()
 
     def close(self):
         self.file.close()


@@ -159,7 +162,7 @@ class Index(IndexStore):
         q = NumericRangeQuery.newIntRange("tag_id", 0, Integer.MAX_VALUE, True, True)
         self.index.deleteDocuments(q)
 
         q = NumericRangeQuery.newIntRange("tag_id", 0, Integer.MAX_VALUE, True, True)
         self.index.deleteDocuments(q)
 

-        for tag in models.Tag.objects.all():
+        for tag in catalogue.models.Tag.objects.all():

             doc = Document()
             doc.add(NumericField("tag_id", Field.Store.YES, True).setIntValue(tag.id))
             doc.add(Field("tag_name", tag.name, Field.Store.NO, Field.Index.ANALYZED))
             doc = Document()
             doc.add(NumericField("tag_id", Field.Store.YES, True).setIntValue(tag.id))
             doc.add(Field("tag_name", tag.name, Field.Store.NO, Field.Index.ANALYZED))


@@ -171,23 +174,23 @@ class Index(IndexStore):
         q = NumericRangeQuery.newIntRange("book_id", book.id, book.id, True, True)
         self.index.deleteDocuments(q)
 
         q = NumericRangeQuery.newIntRange("book_id", book.id, book.id, True, True)
         self.index.deleteDocuments(q)
 

-    def index_book(self, book, overwrite=True):
+    def index_book(self, book, book_info=None, overwrite=True):

         if overwrite:
             self.remove_book(book)
 
         book_doc = self.create_book_doc(book)
         if overwrite:
             self.remove_book(book)
 
         book_doc = self.create_book_doc(book)

-        meta_fields = self.extract_metadata(book)
+        meta_fields = self.extract_metadata(book, book_info)

         for f in meta_fields.values():
         for f in meta_fields.values():

-            if isinstance(f, list):
+            if isinstance(f, list) or isinstance(f, tuple):

                 for elem in f:
                     book_doc.add(elem)
             else:
                 book_doc.add(f)
                 for elem in f:
                     book_doc.add(elem)
             else:
                 book_doc.add(f)

-                
+

         self.index.addDocument(book_doc)
         del book_doc
 
         self.index.addDocument(book_doc)
         del book_doc
 

-        self.index_content(book, book_fields=[meta_fields['title'], meta_fields['author']])
+        self.index_content(book, book_fields=[meta_fields['title'], meta_fields['authors']])

 
     master_tags = [
         'opowiadanie',
 
     master_tags = [
         'opowiadanie',


@@ -210,14 +213,14 @@ class Index(IndexStore):
             doc.add(NumericField("parent_id", Field.Store.YES, True).setIntValue(book.parent.id))
         return doc
 
             doc.add(NumericField("parent_id", Field.Store.YES, True).setIntValue(book.parent.id))
         return doc
 

-    def extract_metadata(self, book):
+    def extract_metadata(self, book, book_info=None):

         fields = {}
         fields = {}

-        book_info = dcparser.parse(book.xml_file)

 
 

-        print("extract metadata for book %s id=%d, thread%d" % (book.slug, book.id, current_thread().ident))
+        if book_info is None:
+            book_info = dcparser.parse(open(book.xml_file.path))

 
         fields['slug'] = Field("slug", book.slug, Field.Store.NO, Field.Index.ANALYZED_NO_NORMS)
 
         fields['slug'] = Field("slug", book.slug, Field.Store.NO, Field.Index.ANALYZED_NO_NORMS)

-        fields['tags'] = [Field("tags", t.name, Field.Store.NO, Field.Index.ANALYZED) for t in book.tags]
+        fields['tags'] = self.add_gaps([Field("tags", t.name, Field.Store.NO, Field.Index.ANALYZED) for t in book.tags], 'tags')

         fields['is_book'] = Field("is_book", 'true', Field.Store.NO, Field.Index.NOT_ANALYZED)
 
         # validator, name
         fields['is_book'] = Field("is_book", 'true', Field.Store.NO, Field.Index.NOT_ANALYZED)
 
         # validator, name


@@ -246,6 +249,7 @@ class Index(IndexStore):
                     dt = getattr(book_info, field.name)
                     fields[field.name] = Field(field.name, "%04d%02d%02d" %\
                                                (dt.year, dt.month, dt.day), Field.Store.NO, Field.Index.NOT_ANALYZED)
                     dt = getattr(book_info, field.name)
                     fields[field.name] = Field(field.name, "%04d%02d%02d" %\
                                                (dt.year, dt.month, dt.day), Field.Store.NO, Field.Index.NOT_ANALYZED)

+

         return fields
 
     def get_master(self, root):
         return fields
 
     def get_master(self, root):


@@ -253,8 +257,14 @@ class Index(IndexStore):
             if master.tag in self.master_tags:
                 return master
 
             if master.tag in self.master_tags:
                 return master
 

+    def add_gaps(self, fields, fieldname):
+        def gap():
+            while True:
+                yield Field(fieldname, ' ', Field.Store.NO, Field.Index.NOT_ANALYZED)
+        return reduce(lambda a, b: a + b, zip(fields, gap()))[0:-1]
+

     def index_content(self, book, book_fields=[]):
     def index_content(self, book, book_fields=[]):

-        wld = WLDocument.from_file(book.xml_file.path)
+        wld = WLDocument.from_file(book.xml_file.path, parse_dublincore=False)

         root = wld.edoc.getroot()
 
         master = self.get_master(root)
         root = wld.edoc.getroot()
 
         master = self.get_master(root)


@@ -270,7 +280,7 @@ class Index(IndexStore):
             return
 
         def fix_format(text):
             return
 
         def fix_format(text):

-            return re.sub("/$", "", text, flags=re.M)
+            return re.sub("(?m)/$", "", text)

 
         def add_part(snippets, **fields):
             doc = self.create_book_doc(book)
 
         def add_part(snippets, **fields):
             doc = self.create_book_doc(book)


@@ -285,21 +295,37 @@ class Index(IndexStore):
             doc.add(Field('content', fields["content"], Field.Store.NO, Field.Index.ANALYZED, \
                           Field.TermVector.WITH_POSITIONS_OFFSETS))
 
             doc.add(Field('content', fields["content"], Field.Store.NO, Field.Index.ANALYZED, \
                           Field.TermVector.WITH_POSITIONS_OFFSETS))
 

-            snip_pos = snippets.add(content)
+            snip_pos = snippets.add(fields["content"])

             doc.add(NumericField("snippets_position", Field.Store.YES, True).setIntValue(snip_pos[0]))
             doc.add(NumericField("snippets_length", Field.Store.YES, True).setIntValue(snip_pos[1]))
 
             if 'fragment_anchor' in fields:
                 doc.add(Field("fragment_anchor", fields['fragment_anchor'],
                               Field.Store.YES, Field.Index.NOT_ANALYZED))
             doc.add(NumericField("snippets_position", Field.Store.YES, True).setIntValue(snip_pos[0]))
             doc.add(NumericField("snippets_length", Field.Store.YES, True).setIntValue(snip_pos[1]))
 
             if 'fragment_anchor' in fields:
                 doc.add(Field("fragment_anchor", fields['fragment_anchor'],
                               Field.Store.YES, Field.Index.NOT_ANALYZED))

-                
+

             if 'themes' in fields:
             if 'themes' in fields:

-                for theme in fields['themes']:
-                    doc.add(Field("themes", theme, Field.Store.YES, Field.Index.ANALYZED, Field.TermVector.WITH_POSITIONS))
-                    doc.add(Field("themes_pl", theme, Field.Store.NO, Field.Index.ANALYZED, Field.TermVector.WITH_POSITIONS))
+                themes, themes_pl = zip(*[
+                    (Field("themes", theme, Field.Store.YES, Field.Index.ANALYZED, Field.TermVector.WITH_POSITIONS),
+                     Field("themes_pl", theme, Field.Store.NO, Field.Index.ANALYZED, Field.TermVector.WITH_POSITIONS))
+                     for theme in fields['themes']])
+
+                themes = self.add_gaps(themes, 'themes')
+                themes_pl = self.add_gaps(themes_pl, 'themes_pl')
+
+                for t in themes:
+                    doc.add(t)
+                for t in themes_pl:
+                    doc.add(t)

 
             return doc
 
 
             return doc
 

+        def give_me_utf8(s):
+            if isinstance(s, unicode):
+                return s.encode('utf-8')
+            else:
+                return s
+
+

         fragments = {}
         snippets = Snippets(book.id).open('w')
         try:
         fragments = {}
         snippets = Snippets(book.id).open('w')
         try:


@@ -315,20 +341,23 @@ class Index(IndexStore):
 
                 self.index.addDocument(doc)
 
 
                 self.index.addDocument(doc)
 

-                for start, end in walker(master):
+                for start, end in walker(header):

                     if start is not None and start.tag == 'begin':
                         fid = start.attrib['id'][1:]
                         fragments[fid] = {'content': [], 'themes': [], 'start_section': position, 'start_header': header.tag}
                         fragments[fid]['content'].append(start.tail)
                     elif start is not None and start.tag == 'motyw':
                         fid = start.attrib['id'][1:]
                     if start is not None and start.tag == 'begin':
                         fid = start.attrib['id'][1:]
                         fragments[fid] = {'content': [], 'themes': [], 'start_section': position, 'start_header': header.tag}
                         fragments[fid]['content'].append(start.tail)
                     elif start is not None and start.tag == 'motyw':
                         fid = start.attrib['id'][1:]

-                        fragments[fid]['themes'].append(start.text)
+                        if start.text is not None:
+                            fragments[fid]['themes'] += map(str.strip, map(give_me_utf8, start.text.split(',')))

                         fragments[fid]['content'].append(start.tail)
                     elif start is not None and start.tag == 'end':
                         fid = start.attrib['id'][1:]
                         if fid not in fragments:
                             continue  # a broken <end> node, skip it
                         frag = fragments[fid]
                         fragments[fid]['content'].append(start.tail)
                     elif start is not None and start.tag == 'end':
                         fid = start.attrib['id'][1:]
                         if fid not in fragments:
                             continue  # a broken <end> node, skip it
                         frag = fragments[fid]

+                        if frag['themes'] == []:
+                            continue  # empty themes list.

                         del fragments[fid]
 
                         def jstr(l):
                         del fragments[fid]
 
                         def jstr(l):


@@ -383,35 +412,23 @@ class ReusableIndex(Index):
     if you cannot rely on atexit, use ReusableIndex.close_reusable() yourself.
     """
     index = None
     if you cannot rely on atexit, use ReusableIndex.close_reusable() yourself.
     """
     index = None

-    pool = None
-    pool_jobs = None

 
     def open(self, analyzer=None, threads=4):
         if ReusableIndex.index is not None:
             self.index = ReusableIndex.index
         else:
             print("opening index")
 
     def open(self, analyzer=None, threads=4):
         if ReusableIndex.index is not None:
             self.index = ReusableIndex.index
         else:
             print("opening index")

-            ReusableIndex.pool = ThreadPool(threads, initializer=lambda: JVM.attachCurrentThread() )
-            ReusableIndex.pool_jobs = []

             Index.open(self, analyzer)
             ReusableIndex.index = self.index
             atexit.register(ReusableIndex.close_reusable)
 
             Index.open(self, analyzer)
             ReusableIndex.index = self.index
             atexit.register(ReusableIndex.close_reusable)
 

-    def index_book(self, *args, **kw):
-        job = ReusableIndex.pool.apply_async(log_exception_wrapper(Index.index_book), (self,) + args, kw)
-        ReusableIndex.pool_jobs.append(job)
+    # def index_book(self, *args, **kw):
+    #     job = ReusableIndex.pool.apply_async(log_exception_wrapper(Index.index_book), (self,) + args, kw)
+    #     ReusableIndex.pool_jobs.append(job)

 
     @staticmethod
     def close_reusable():
         if ReusableIndex.index is not None:
 
     @staticmethod
     def close_reusable():
         if ReusableIndex.index is not None:

-            print("wait for indexing to finish")
-            for job in ReusableIndex.pool_jobs:
-                job.get()
-                sys.stdout.write('.')
-                sys.stdout.flush()
-            print("done.")
-            ReusableIndex.pool.close()
-

             ReusableIndex.index.optimize()
             ReusableIndex.index.close()
             ReusableIndex.index = None
             ReusableIndex.index.optimize()
             ReusableIndex.index.close()
             ReusableIndex.index = None


@@ -466,9 +483,10 @@ class Search(IndexStore):
         bks = []
         for found in tops.scoreDocs:
             doc = self.searcher.doc(found.doc)
         bks = []
         for found in tops.scoreDocs:
             doc = self.searcher.doc(found.doc)

-            bks.append(models.Book.objects.get(id=doc.get("book_id")))
+            bks.append(catalogue.models.Book.objects.get(id=doc.get("book_id")))

         return (bks, tops.totalHits)
 
         return (bks, tops.totalHits)
 

+

     def search(self, query, max_results=50):
         query = self.query(query)
         query = self.wrapjoins(query, ["content", "themes"])
     def search(self, query, max_results=50):
         query = self.query(query)
         query = self.wrapjoins(query, ["content", "themes"])


@@ -477,7 +495,7 @@ class Search(IndexStore):
         bks = []
         for found in tops.scoreDocs:
             doc = self.searcher.doc(found.doc)
         bks = []
         for found in tops.scoreDocs:
             doc = self.searcher.doc(found.doc)

-            bks.append(models.Book.objects.get(id=doc.get("book_id")))
+            bks.append(catalogue.models.Book.objects.get(id=doc.get("book_id")))

         return (bks, tops.totalHits)
 
     def bsearch(self, query, max_results=50):
         return (bks, tops.totalHits)
 
     def bsearch(self, query, max_results=50):


@@ -488,7 +506,7 @@ class Search(IndexStore):
         bks = []
         for found in tops.scoreDocs:
             doc = self.searcher.doc(found.doc)
         bks = []
         for found in tops.scoreDocs:
             doc = self.searcher.doc(found.doc)

-            bks.append(models.Book.objects.get(id=doc.get("book_id")))
+            bks.append(catalogue.models.Book.objects.get(id=doc.get("book_id")))

         return (bks, tops.totalHits)
 
 # TokenStream tokenStream = analyzer.tokenStream(fieldName, reader);
         return (bks, tops.totalHits)
 
 # TokenStream tokenStream = analyzer.tokenStream(fieldName, reader);


@@ -503,7 +521,9 @@ class Search(IndexStore):
 
 
 class SearchResult(object):
 
 
 class SearchResult(object):

-    def __init__(self, searcher, scoreDocs, score=None, how_found=None, snippets_cb=None):
+    def __init__(self, searcher, scoreDocs, score=None, how_found=None, snippets=None):
+        self.snippets = []
+

         if score:
             self.score = score
         else:
         if score:
             self.score = score
         else:


@@ -515,15 +535,16 @@ class SearchResult(object):
         self.book_id = int(stored.get("book_id"))
 
         header_type = stored.get("header_type")
         self.book_id = int(stored.get("book_id"))
 
         header_type = stored.get("header_type")

+        if not header_type:
+            return
+

         sec = (header_type, int(stored.get("header_index")))
         header_span = stored.get('header_span')
         header_span = header_span is not None and int(header_span) or 1
         sec = (header_type, int(stored.get("header_index")))
         header_span = stored.get('header_span')
         header_span = header_span is not None and int(header_span) or 1

-        stored = searcher.doc(scoreDocs.doc)
-        self.book_id = int(stored.get("book_id"))

 
         fragment = stored.get("fragment_anchor")
 
 
         fragment = stored.get("fragment_anchor")
 

-        hit = (sec + (header_span,), fragment, scoreDocs.score, {'how_found': how_found, 'snippets_cb': snippets_cb})
+        hit = (sec + (header_span,), fragment, scoreDocs.score, {'how_found': how_found, 'snippets': snippets})

 
         self.hits.append(hit)
 
 
         self.hits.append(hit)
 


@@ -535,37 +556,41 @@ class SearchResult(object):
             self.score = other.score
         return self
 
             self.score = other.score
         return self
 

-    def add_snippets(self, snippets):
-        self.snippets += snippets
-        return self
-

     def get_book(self):
     def get_book(self):

-        return models.Book.objects.get(id=self.book_id)
+        return catalogue.models.Book.objects.get(id=self.book_id)

 
     book = property(get_book)
 
 
     book = property(get_book)
 

-    def get_parts(self):
-        book = self.book
+    def process_hits(self):
+        frags = filter(lambda r: r[1] is not None, self.hits)
+        sect = filter(lambda r: r[1] is None, self.hits)
+        sect = filter(lambda s: 0 == len(filter(
+            lambda f: s[0][1] >= f[0][1] and s[0][1] < f[0][1] + f[0][2],
+            frags)), sect)
+
+        hits = []

 
 

-        def sections_covered(results):
-            frags = filter(lambda r: r[1] is not None, results)
-            sect = filter(lambda r: r[1] is None, results)
-            sect = filter(lambda s: 0 == len(filter(
-                lambda f: s[0][1] >= f[0][1] and s[0][1] < f[0][1] + f[0][2],
-                frags)), sect)
-            print "filtered, non overlapped sections: %s" % sect
-            return frags + sect
-            
-            
-        parts = [{"header": s[0], "position": s[1], '_score_key': s} for s in self.sections] \
-            + [{"fragment": book.fragments.get(anchor=f), '_score_key':f} for f in self.fragments]
+        for s in sect:
+            m = {'score': s[2],
+                 'header_index': s[0][1]
+                 }
+            m.update(s[3])
+            hits.append(m)

 
 

-        parts.sort(lambda a, b: cmp(self.scores[a['_score_key']], self.scores[b['_score_key']]))
-        print("bookid: %d parts: %s" % (self.book_id, parts))
-        return parts
+        for f in frags:
+            frag = catalogue.models.Fragment.objects.get(anchor=f[1])
+            m = {'score': f[2],
+                 'fragment': frag,
+                 'themes': frag.tags.filter(category='theme')
+                 }
+            m.update(f[3])
+            hits.append(m)

 
 

-    parts = property(get_parts)
+        hits.sort(lambda a, b: cmp(a['score'], b['score']), reverse=True)

 
 

+        print("--- %s" % hits)
+
+        return hits

 
     def __unicode__(self):
         return u'SearchResult(book_id=%d, score=%d)' % (self.book_id, self.score)
 
     def __unicode__(self):
         return u'SearchResult(book_id=%d, score=%d)' % (self.book_id, self.score)


@@ -586,6 +611,75 @@ class SearchResult(object):
         return cmp(self.score, other.score)
 
 
         return cmp(self.score, other.score)
 
 

+class Hint(object):
+    def __init__(self, search):
+        self.search = search
+        self.book_tags = {}
+        self.part_tags = []
+        self._books = []
+
+    def books(self, *books):
+        self._books = books
+
+    def tags(self, tags):
+        for t in tags:
+            if t.category in ['author', 'title', 'epoch', 'genre', 'kind']:
+                lst = self.book_tags.get(t.category, [])
+                lst.append(t)
+                self.book_tags[t.category] = lst
+            if t.category in ['theme']:
+                self.part_tags.append(t)
+
+    def tag_filter(self, tags, field='tags'):
+        q = BooleanQuery()
+
+        for tag in tags:
+            toks = self.search.get_tokens(tag.name, field=field)
+            tag_phrase = PhraseQuery()
+            for tok in toks:
+                tag_phrase.add(Term(field, tok))
+            q.add(BooleanClause(tag_phrase, BooleanClause.Occur.MUST))
+
+        return QueryWrapperFilter(q)
+
+    def book_filter(self):
+        tags = reduce(lambda a, b: a + b, self.book_tags.values(), [])
+        if tags:
+            return self.tag_filter(tags)
+        else:
+            return None
+
+    def part_filter(self):
+        fs = []
+        if self.part_tags:
+            fs.append(self.tag_filter(self.part_tags, field='themes'))
+
+        if self._books != []:
+            bf = BooleanFilter()
+            for b in self._books:
+                id_filter = NumericRangeFilter.newIntRange('book_id', b.id, b.id, True, True)
+                bf.add(FilterClause(id_filter, BooleanClause.Occur.SHOULD))
+            fs.append(bf)
+
+        return MultiSearch.chain_filters(fs)
+
+    def should_search_for_book(self):
+        return self._books == []
+
+    def just_search_in(self, all):
+        """Holds logic to figure out which indexes should be search, when we have some hinst already"""
+        some = []
+        for field in all:
+            if field == 'authors' and 'author' in self.book_tags:
+                continue
+            if field == 'title' and self._books != []:
+                continue
+            if (field == 'themes' or field == 'themes_pl') and self.part_tags:
+                continue
+            some.append(field)
+        return some
+
+

 class MultiSearch(Search):
     """Class capable of IMDb-like searching"""
     def get_tokens(self, searched, field='content'):
 class MultiSearch(Search):
     """Class capable of IMDb-like searching"""
     def get_tokens(self, searched, field='content'):


@@ -651,72 +745,83 @@ class MultiSearch(Search):
             q.add(BooleanClause(term, modal))
         return q
 
             q.add(BooleanClause(term, modal))
         return q
 

-    def content_query(self, query):
-        return BlockJoinQuery(query, self.parent_filter,
-                              BlockJoinQuery.ScoreMode.Total)
+    # def content_query(self, query):
+    #     return BlockJoinQuery(query, self.parent_filter,
+    #                           BlockJoinQuery.ScoreMode.Total)

 
 

-    def search_perfect_book(self, searched, max_results=20, fuzzy=False):
-        qrys = [self.make_phrase(self.get_tokens(searched, field=fld), field=fld, fuzzy=fuzzy) for fld in ['author', 'title']]
+    def search_perfect_book(self, searched, max_results=20, fuzzy=False, hint=None):
+        fields_to_search = ['authors', 'title']
+        only_in = None
+        if hint:
+            if not hint.should_search_for_book():
+                return []
+            fields_to_search = hint.just_search_in(fields_to_search)
+            only_in = hint.book_filter()
+
+        qrys = [self.make_phrase(self.get_tokens(searched, field=fld), field=fld, fuzzy=fuzzy) for fld in fields_to_search]

 
         books = []
         for q in qrys:
 
         books = []
         for q in qrys:

-            top = self.searcher.search(q, max_results)
+            top = self.searcher.search(q,
+                                       self.chain_filters([only_in, self.term_filter(Term('is_book', 'true'))]),
+                max_results)

             for found in top.scoreDocs:
                 books.append(SearchResult(self.searcher, found))
         return books
 
             for found in top.scoreDocs:
                 books.append(SearchResult(self.searcher, found))
         return books
 

-    def search_perfect_parts(self, searched, max_results=20, fuzzy=False):
+    def search_perfect_parts(self, searched, max_results=20, fuzzy=False, hint=None):

         qrys = [self.make_phrase(self.get_tokens(searched), field=fld, fuzzy=fuzzy) for fld in ['content']]
 
         qrys = [self.make_phrase(self.get_tokens(searched), field=fld, fuzzy=fuzzy) for fld in ['content']]
 

+        flt = None
+        if hint:
+            flt = hint.part_filter()
+

         books = []
         for q in qrys:
         books = []
         for q in qrys:

-            top = self.searcher.search(q, max_results)
+            top = self.searcher.search(q,
+                                       self.chain_filters([self.term_filter(Term('is_book', 'true'), inverse=True),
+                                                           flt
+                                                          ]),
+                                       max_results)

             for found in top.scoreDocs:
             for found in top.scoreDocs:

-                books.append(SearchResult(self.searcher, found).add_snippets(self.get_snippets(found, q)))
+                books.append(SearchResult(self.searcher, found, snippets=self.get_snippets(found, q)))

 
         return books
 
 
         return books
 

-    def search_everywhere(self, searched, max_results=20, fuzzy=False):
+    def search_everywhere(self, searched, max_results=20, fuzzy=False, hint=None):

         books = []
         books = []

+        only_in = None
+
+        if hint:
+            only_in = hint.part_filter()

 
         # content only query : themes x content
         q = BooleanQuery()
 
         tokens = self.get_tokens(searched)
 
         # content only query : themes x content
         q = BooleanQuery()
 
         tokens = self.get_tokens(searched)

-        q.add(BooleanClause(self.make_term_query(tokens, field='themes', fuzzy=fuzzy), BooleanClause.Occur.MUST))
-        q.add(BooleanClause(self.make_term_query(tokens, field='content', fuzzy=fuzzy), BooleanClause.Occur.SHOULD))
+        if hint is None or hint.just_search_in(['themes_pl']) != []:
+            q.add(BooleanClause(self.make_term_query(tokens, field='themes_pl',
+                                                     fuzzy=fuzzy), BooleanClause.Occur.MUST))
+
+        q.add(BooleanClause(self.make_term_query(tokens, field='content',
+                                                 fuzzy=fuzzy), BooleanClause.Occur.SHOULD))

 
 

-        topDocs = self.searcher.search(q, max_results)
+        topDocs = self.searcher.search(q, only_in, max_results)

         for found in topDocs.scoreDocs:
             books.append(SearchResult(self.searcher, found))
 
         for found in topDocs.scoreDocs:
             books.append(SearchResult(self.searcher, found))
 

-        # joined query themes/content x author/title/epochs/genres/kinds
-        # q = BooleanQuery()
-        # in_meta = BooleanQuery()
-        # in_content = BooleanQuery()
-
-        # for fld in ['themes', 'content']:
-        #     in_content.add(BooleanClause(self.make_term_query(tokens, field=fld, fuzzy=False), BooleanClause.Occur.SHOULD))
-
-        # in_meta.add(BooleanClause(self.make_term_query(
-        #     self.get_tokens(searched, field='author'), field='author', fuzzy=fuzzy), BooleanClause.Occur.SHOULD))
-
-        # for fld in ['title', 'epochs', 'genres', 'kinds']:
-        #     in_meta.add(BooleanClause(self.make_term_query(tokens, field=fld, fuzzy=fuzzy), BooleanClause.Occur.SHOULD))
+        # query themes/content x author/title/tags
+        q = BooleanQuery()
+        #        in_meta = BooleanQuery()
+        in_content = BooleanQuery()

 
 

-        # q.add(BooleanClause(in_meta, BooleanClause.Occur.MUST))
-        # in_content_join = self.content_query(in_content)
-        # q.add(BooleanClause(in_content_join, BooleanClause.Occur.MUST))
-        # #        import pdb; pdb.set_trace()
-        # collector = BlockJoinCollector(Sort.RELEVANCE, 100, True, True)
+        for fld in ['themes', 'content', 'tags', 'authors', 'title']:
+            in_content.add(BooleanClause(self.make_term_query(tokens, field=fld, fuzzy=False), BooleanClause.Occur.SHOULD))

 
 

-        # self.searcher.search(q, collector)
+        topDocs = self.searcher.search(q, only_in, max_results)
+        for found in topDocs.scoreDocs:
+            books.append(SearchResult(self.searcher, found))

 
 

-        # top_groups = collector.getTopGroups(in_content_join, Sort.RELEVANCE, 0, max_results, 0, True)
-        # if top_groups:
-        #     for grp in top_groups.groups:
-        #         for part in grp.scoreDocs:
-        #             books.append(SearchResult(self.searcher, part, score=grp.maxScore))

         return books
 
     def multisearch(self, query, max_results=50):
         return books
 
     def multisearch(self, query, max_results=50):


@@ -724,8 +829,8 @@ class MultiSearch(Search):
         Search strategy:
         - (phrase) OR -> content
                       -> title
         Search strategy:
         - (phrase) OR -> content
                       -> title

-                      -> author
-        - (keywords)  -> author
+                      -> authors
+        - (keywords)  -> authors

                       -> motyw
                       -> tags
                       -> content
                       -> motyw
                       -> tags
                       -> content


@@ -768,7 +873,7 @@ class MultiSearch(Search):
         bks = []
         for found in tops.scoreDocs:
             doc = self.searcher.doc(found.doc)
         bks = []
         for found in tops.scoreDocs:
             doc = self.searcher.doc(found.doc)

-            b = models.Book.objects.get(id=doc.get("book_id"))
+            b = catalogue.models.Book.objects.get(id=doc.get("book_id"))

             bks.append(b)
             print "%s (%d) -> %f" % (b, b.id, found.score)
         return bks
             bks.append(b)
             print "%s (%d) -> %f" % (b, b.id, found.score)
         return bks


@@ -782,14 +887,15 @@ class MultiSearch(Search):
         # locate content.
         snippets = Snippets(stored.get('book_id')).open()
         try:
         # locate content.
         snippets = Snippets(stored.get('book_id')).open()
         try:

-            text = snippets.get(stored.get('snippets_position'), stored.get('snippets_length'))
+            text = snippets.get((int(stored.get('snippets_position')),
+                                 int(stored.get('snippets_length'))))

         finally:
             snippets.close()
 
         tokenStream = TokenSources.getAnyTokenStream(self.searcher.getIndexReader(), scoreDoc.doc, field, self.analyzer)
         #  highlighter.getBestTextFragments(tokenStream, text, False, 10)
         finally:
             snippets.close()
 
         tokenStream = TokenSources.getAnyTokenStream(self.searcher.getIndexReader(), scoreDoc.doc, field, self.analyzer)
         #  highlighter.getBestTextFragments(tokenStream, text, False, 10)

+        #        import pdb; pdb.set_trace()

         snip = highlighter.getBestFragments(tokenStream, text, 3, "...")
         snip = highlighter.getBestFragments(tokenStream, text, 3, "...")

-        print('snips: %s' % snip)

 
         return [snip]
 
 
         return [snip]
 


@@ -816,12 +922,20 @@ class MultiSearch(Search):
         tags = []
         for found in tops.scoreDocs:
             doc = self.searcher.doc(found.doc)
         tags = []
         for found in tops.scoreDocs:
             doc = self.searcher.doc(found.doc)

-            tag = models.Tag.objects.get(id=doc.get("tag_id"))
+            tag = catalogue.models.Tag.objects.get(id=doc.get("tag_id"))

             tags.append(tag)
             print "%s (%d) -> %f" % (tag, tag.id, found.score)
 
         return tags
 
             tags.append(tag)
             print "%s (%d) -> %f" % (tag, tag.id, found.score)
 
         return tags
 

+    def search_books(self, query, filter=None, max_results=10):
+        bks = []
+        tops = self.searcher.search(query, filter, max_results)
+        for found in tops.scoreDocs:
+            doc = self.searcher.doc(found.doc)
+            bks.append(catalogue.models.Book.objects.get(id=doc.get("book_id")))
+        return bks
+

     def create_prefix_phrase(self, toks, field):
         q = MultiPhraseQuery()
         for i in range(len(toks)):
     def create_prefix_phrase(self, toks, field):
         q = MultiPhraseQuery()
         for i in range(len(toks)):


@@ -836,6 +950,18 @@ class MultiSearch(Search):
                 q.add(t)
         return q
 
                 q.add(t)
         return q
 

+    @staticmethod
+    def term_filter(term, inverse=False):
+        only_term = TermsFilter()
+        only_term.addTerm(term)
+
+        if inverse:
+            neg = BooleanFilter()
+            neg.add(FilterClause(only_term, BooleanClause.Occur.MUST_NOT))
+            only_term = neg
+
+        return only_term
+

     def hint_tags(self, string, max_results=50):
         toks = self.get_tokens(string, field='SIMPLE')
         top = BooleanQuery()
     def hint_tags(self, string, max_results=50):
         toks = self.get_tokens(string, field='SIMPLE')
         top = BooleanQuery()


@@ -844,11 +970,7 @@ class MultiSearch(Search):
             q = self.create_prefix_phrase(toks, field)
             top.add(BooleanClause(q, BooleanClause.Occur.SHOULD))
 
             q = self.create_prefix_phrase(toks, field)
             top.add(BooleanClause(q, BooleanClause.Occur.SHOULD))
 

-        book_cat = TermsFilter()
-        book_cat.addTerm(Term("tag_category", "book"))
-
-        no_book_cat = BooleanFilter()
-        no_book_cat.add(FilterClause(book_cat, BooleanClause.Occur.MUST_NOT))
+        no_book_cat = self.term_filter(Term("tag_category", "book"), inverse=True)

 
         return self.search_tags(top, no_book_cat, max_results=max_results)
 
 
         return self.search_tags(top, no_book_cat, max_results=max_results)
 


@@ -856,25 +978,22 @@ class MultiSearch(Search):
         toks = self.get_tokens(string, field='SIMPLE')
 
         q = self.create_prefix_phrase(toks, 'title')
         toks = self.get_tokens(string, field='SIMPLE')
 
         q = self.create_prefix_phrase(toks, 'title')

-        only_books = TermsFilter()
-        only_books.addTerm(Term("is_book", "true"))
-
-        return self.book_search(q, only_books, max_results=max_results)

 
 

-    def filter_by_tags(self, tags):
-        q = BooleanQuery()
-        
-        for tag in tags:
-            toks = self.get_tokens(tag.name, field='tags')
-            tag_phrase = PhraseQuery()
-            for tok in toks:
-                tag_phrase.add(tok)
-            q.add(BooleanClause(tok, BooleanClause.Occur.MUST))
+        return self.book_search(q, self.term_filter(Term("is_book", "true")), max_results=max_results)

 
 

-        return QueryWrapperFilter(q)
+    @staticmethod
+    def chain_filters(filters, op=ChainedFilter.AND):
+        filters = filter(lambda x: x is not None, filters)
+        if not filters:
+            return None
+        chf = ChainedFilter(JArray('object')(filters, Filter), op)
+        return chf

 
     def filtered_categories(self, tags):
         cats = {}
         for t in tags:
             cats[t.category] = True
         return cats.keys()
 
     def filtered_categories(self, tags):
         cats = {}
         for t in tags:
             cats[t.category] = True
         return cats.keys()

+
+    def hint(self):
+        return Hint(self)