Searches: manage AND, OR, parentheses and open search '*' (refs #4180)

1 files changed, 111 insertions, 6 deletions

diff --git a/ishtar_common/views_item.py b/ishtar_common/views_item.py
index 513035903..f5e47a832 100644
--- a/ishtar_common/views_item.py
+++ b/ishtar_common/views_item.py
@@ -268,11 +268,107 @@ def _get_values(request, val):
     return new_vals
 
 
+def _push_to_list(obj, current_group, depth):
+    """
+    parse_parentheses helper function
+    """
+    try:
+        while depth > 0:
+            current_group = current_group[-1]
+            depth -= 1
+    except IndexError:
+        # tolerant to parentheses mismatch
+        pass
+    if current_group and type(obj) in (unicode, str) and \
+            type(current_group[-1]) in (unicode, str):
+        current_group[-1] += obj
+    else:
+        current_group.append(obj)
+
+
+def _parse_parentheses(s):
+    """
+    Parse parentheses into list.
+    (OA01 & (pierre | ciseau)) -> ["0A01 &", ["pierre | ciseau"]]
+    """
+
+    groups = []
+    depth = 0
+
+    for char in s:
+        if char == u'(':
+            _push_to_list([], groups, depth)
+            depth += 1
+        elif char == ')':
+            if depth > 0:
+                depth -= 1
+        else:
+            _push_to_list(char, groups, depth)
+    # for non tolerant to parentheses mismatch check depth is equal to 0
+    return groups
+
+
+FORBIDDEN_CHAR = [u":"]
+RESERVED_CHAR = [u"|", u"&"]
+
+
+def _parse_query_string(string):
+    string = string.strip().lower()
+    for reserved_char in FORBIDDEN_CHAR:
+        string = string.replace(reserved_char, u"")
+    if len(string) != 1:
+        for reserved_char in RESERVED_CHAR:
+            string = string.replace(reserved_char, u"")
+    # like search
+    if string.endswith(u'*'):
+        string = string[:-1] + u':*'
+
+    return string
+
+
+def _parse_parentheses_groups(groups):
+    """
+    Transform parentheses groups to query
+    """
+    if type(groups) is not list:
+        string = groups.strip()
+        # split into many groups if spaces
+        if ' ' not in string:
+            return _parse_query_string(groups)
+        return _parse_parentheses_groups(string.split(u" "))
+    if not groups:  # empty list
+        return ""
+    query = u"("
+    previous_sep, has_item = None, False
+    for item in groups:
+        q = _parse_parentheses_groups(item).strip()
+        if not q:
+            continue
+        if q in (u"|", u"&"):
+            if previous_sep or not has_item:
+                continue  # multiple sep is not relevant
+            previous_sep = q
+            continue
+        if has_item:
+            if previous_sep:
+                query += previous_sep
+            else:
+                query += u" & "
+        query += q
+        has_item = True
+        previous_sep = None
+    query += u")"
+    return unidecode(query)
+
+
 def _search_manage_search_vector(dct):
     if 'search_vector' in dct:
-        dct['search_vector'] = SearchQuery(
-            unidecode(dct['search_vector']),
-            config=settings.ISHTAR_SEARCH_LANGUAGE
+        parentheses_groups = _parse_parentheses(dct['search_vector'].strip())
+        query = _parse_parentheses_groups(parentheses_groups)
+        dct['extras'].append(
+            {'where': ["search_vector @@ (to_tsquery(%s, %s)) = true"],
+             'params': [settings.ISHTAR_SEARCH_LANGUAGE,
+                        query]}
         )
     return dct
 
@@ -594,7 +690,13 @@ def get_item(model, func_name, default_name, extra_request_keys=[],
                         reqs |= q
                     and_reqs.append(reqs)
                     break
+        dct['extras'] = []
         dct = _search_manage_search_vector(dct)
+        search_vector = ""
+        if 'search_vector' in dct:
+            search_vector = dct.pop('search_vector')
+        extras = dct.pop('extras')
+
         query = Q(**dct)
         for k, or_req in or_reqs:
             alt_dct = dct.copy()
@@ -665,11 +767,14 @@ def get_item(model, func_name, default_name, extra_request_keys=[],
                     dct = {upper_key: current}
                     query &= Q(**dct)
 
-        items = model.objects.filter(query).distinct()
+        items = model.objects.filter(query)
+        for extra in extras:
+            items = items.extra(**extra)
+        items = items.distinct()
         # print(items.query)
 
-        if 'search_vector' in dct:  # for serialization
-            dct['search_vector'] = dct['search_vector'].value
+        if search_vector:  # for serialization
+            dct['search_vector'] = search_vector
 
         # table cols
         if own_table_cols: