From 80ded6938c72c9bcb0b9ec026fe8c7fcb9bc65cf Mon Sep 17 00:00:00 2001
From: Igor Chubin <igor@chub.in>
Date: Wed, 5 Aug 2020 09:46:36 +0200
Subject: [PATCH] fixed search (#147, #225)

---
 lib/postprocessing.py | 38 ++-------------------------
 lib/search.py         | 61 ++++++++++++++++++++++++++++++++++++++-----
 2 files changed, 57 insertions(+), 42 deletions(-)

diff --git a/lib/postprocessing.py b/lib/postprocessing.py
index d54a359..bc82c1e 100644
--- a/lib/postprocessing.py
+++ b/lib/postprocessing.py
@@ -1,4 +1,4 @@
-import re
+import search
 import fmt.comments
 
 def postprocess(answer, keyword, options, request_options=None):
@@ -49,42 +49,8 @@ def _filter_by_keyword(answer, keyword, options):
         answer.append(paragraph)
         return answer
 
-    def _paragraph_contains(paragraph, keyword, insensitive=False, word_boundaries=True):
-        """
-        Check if `paragraph` contains `keyword`.
-        Several keywords can be joined together using ~
-        For example: ~ssh~passphrase
-        """
-        answer = True
-
-        if '~' in keyword:
-            keywords = keyword.split('~')
-        else:
-            keywords = [keyword]
-
-        for kwrd in keywords:
-            regex = re.escape(kwrd)
-            if not word_boundaries:
-                regex = r"\b%s\b" % kwrd
-
-            if insensitive:
-                answer = answer and bool(re.search(regex, paragraph, re.IGNORECASE))
-            else:
-                answer = answer and bool(re.search(regex, paragraph))
-
-        return answer
-
-
-    if not keyword:
-        return answer
-
-    search_options = {
-        'insensitive': 'i' in options,
-        'word_boundaries': 'b' in options
-    }
-
     paragraphs = [p for p in _split_paragraphs(answer)
-                  if _paragraph_contains(p, keyword, **search_options)]
+                  if search.match(p, keyword, options=options)]
     if not paragraphs:
         return ""
 
diff --git a/lib/search.py b/lib/search.py
index 9b0bfed..58e205c 100644
--- a/lib/search.py
+++ b/lib/search.py
@@ -19,6 +19,8 @@ Configuration parameters:
     search.limit
 """
 
+import re
+
 from config import CONFIG
 from routing import get_answer_dict, get_topics_list
 
@@ -30,13 +32,60 @@ def _limited_entry():
         'format': "code",
     }
 
+def _parse_options(options):
+    """Parse search options string into optiond_dict
+    """
+
+    if options is None:
+        return {}
+
+    search_options = {
+        'insensitive': 'i' in options,
+        'word_boundaries': 'b' in options,
+        'recursive': 'r' in options,
+    }
+    return search_options
+
+def match(paragraph, keyword, options=None, options_dict=None):
+    """Search for each keyword from `keywords` in `page`
+    and if all of them are found, return `True`.
+    Otherwise return `False`.
+
+    Several keywords can be joined together using ~
+    For example: ~ssh~passphrase
+    """
+
+    if '~' in keyword:
+        keywords = keyword.split('~')
+    else:
+        keywords = [keyword]
+
+    if options_dict is None:
+        options_dict = _parse_options(options)
+
+    for kwrd in keywords:
+        if not kwrd:
+            continue
+
+        regex = re.escape(kwrd)
+        if options_dict["word_boundaries"]:
+            regex = r"\b%s\b" % kwrd
+
+        if options_dict["insensitive"]:
+            if not re.search(regex, paragraph, re.IGNORECASE):
+                return False
+        else:
+            if not re.search(regex, paragraph):
+                return False
+    return True
+
 def find_answers_by_keyword(directory, keyword, options="", request_options=None):
     """
     Search in the whole tree of all cheatsheets or in its subtree `directory`
     by `keyword`
     """
 
-    recursive = 'r' in options
+    options_dict = _parse_options(options)
 
     answers_found = []
     for topic in get_topics_list(skip_internal=True, skip_dirs=True):
@@ -45,13 +94,13 @@ def find_answers_by_keyword(directory, keyword, options="", request_options=None
             continue
 
         subtopic = topic[len(directory):]
-        if not recursive and '/' in subtopic:
+        if not options_dict["recursive"] and '/' in subtopic:
             continue
 
-        answer = get_answer_dict(topic, request_options=request_options)
-
-        if answer and answer.get('answer') and keyword.lower() in answer.get('answer', '').lower():
-            answers_found.append(answer)
+        answer_dict = get_answer_dict(topic, request_options=request_options)
+        answer_text = answer_dict.get('answer', '')
+        if match(answer_text, keyword, options_dict=options_dict):
+            answers_found.append(answer_dict)
 
         if len(answers_found) > CONFIG['search.limit']:
             answers_found.append(