Update README.rst

Clarified that some signature extraction methods require initializing the lib first.

README.rst

@@ -3,7 +3,7 @@ talon
 
 Mailgun library to extract message quotations and signatures.
 
-If you ever tried to parse message quotations or signatures you know that absense of any formatting standards in this area could make this task a nightmare. Hopefully this library will make your life much easier. The name of the project is inspired by TALON - multipurpose robot designed to perform missions ranging from reconnaissance to combat and operate in a number of hostile environments. That’s what a good quotations and signature parser should be like :smile:
+If you ever tried to parse message quotations or signatures you know that absence of any formatting standards in this area could make this task a nightmare. Hopefully this library will make your life much easier. The name of the project is inspired by TALON - multipurpose robot designed to perform missions ranging from reconnaissance to combat and operate in a number of hostile environments. That’s what a good quotations and signature parser should be like :smile:
 
 Usage
 -----
@@ -71,6 +71,11 @@ the power of machine learning algorithms:
 
 .. code:: python
 
+    import talon
+    # don't forget to init the library first
+    # it loads machine learning classifiers
+    talon.init()
+
     from talon import signature
 
 

setup.py

@@ -26,7 +26,8 @@ setup(name='talon',
           "html2text",
           "nose==1.2.1",
           "mock",
-          "coverage"
+          "coverage",
+          "flanker"
           ]
       )
 

talon/quotations.py

@@ -73,6 +73,9 @@ SPLITTER_PATTERNS = [
     re.compile("(\d+/\d+/\d+|\d+\.\d+\.\d+).*@", re.VERBOSE),
     RE_ON_DATE_SMB_WROTE,
     re.compile('(_+\r?\n)?[\s]*(:?[*]?From|Date):[*]? .*'),
+    re.compile('(_+\r?\n)?[\s]*(:?[*]?Van|Datum):[*]? .*'),
+    re.compile('(_+\r?\n)?[\s]*(:?[*]?De|Date):[*]? .*'),
+    re.compile('(_+\r?\n)?[\s]*(:?[*]?Von|Datum):[*]? .*'),
     re.compile('\S{3,10}, \d\d? \S{3,10} 20\d\d,? \d\d?:\d\d(:\d\d)?'
                '( \S+){3,6}@\S+:')
     ]
@@ -81,7 +84,7 @@ SPLITTER_PATTERNS = [
 RE_LINK = re.compile('<(http://[^>]*)>')
 RE_NORMALIZED_LINK = re.compile('@@(http://[^>@]*)@@')
 
-RE_PARANTHESIS_LINK = re.compile("\(https?://")
+RE_PARENTHESIS_LINK = re.compile("\(https?://")
 
 SPLITTER_MAX_LINES = 4
 MAX_LINES_COUNT = 1000
@@ -169,8 +172,8 @@ def process_marked_lines(lines, markers, return_flags=[False, -1, -1]):
         # long links could break sequence of quotation lines but they shouldn't
         # be considered an inline reply
         links = (
-            RE_PARANTHESIS_LINK.search(lines[inline_reply.start() - 1]) or
+            RE_PARENTHESIS_LINK.search(lines[inline_reply.start() - 1]) or
-            RE_PARANTHESIS_LINK.match(lines[inline_reply.start()].strip()))
+            RE_PARENTHESIS_LINK.match(lines[inline_reply.start()].strip()))
         if not links:
             return_flags[:] = [False, -1, -1]
             return lines
@@ -197,7 +200,7 @@ def preprocess(msg_body, delimiter, content_type='text/plain'):
     """Prepares msg_body for being stripped.
 
     Replaces link brackets so that they couldn't be taken for quotation marker.
-    Splits line in two if splitter pattern preceeded by some text on the same
+    Splits line in two if splitter pattern preceded by some text on the same
     line (done only for 'On <date> <person> wrote:' pattern).
     """
     # normalize links i.e. replace '<', '>' wrapping the link with some symbols
@@ -213,7 +216,7 @@ def preprocess(msg_body, delimiter, content_type='text/plain'):
     msg_body = re.sub(RE_LINK, link_wrapper, msg_body)
 
     def splitter_wrapper(splitter):
-        """Wrapps splitter with new line"""
+        """Wraps splitter with new line"""
         if splitter.start() and msg_body[splitter.start() - 1] != '\n':
             return '%s%s' % (delimiter, splitter.group())
         else:
@@ -268,7 +271,7 @@ def extract_from_html(msg_body):
     then converting html to text,
     then extracting quotations from text,
     then checking deleted checkpoints,
-    then deleting neccessary tags.
+    then deleting necessary tags.
     """
 
     if msg_body.strip() == '':

talon/signature/bruteforce.py

@@ -49,7 +49,7 @@ RE_PHONE_SIGNATURE = re.compile(r'''
 # c - could be signature line
 # d - line starts with dashes (could be signature or list item)
 # l - long line
-RE_SIGNATURE_CANDIDAATE = re.compile(r'''
+RE_SIGNATURE_CANDIDATE = re.compile(r'''
     (?P<candidate>c+d)[^d]
     |
     (?P<candidate>c+d)$
@@ -184,5 +184,5 @@ def _process_marked_candidate_indexes(candidate, markers):
     >>> _process_marked_candidate_indexes([9, 12, 14, 15, 17], 'clddc')
     [15, 17]
     """
-    match = RE_SIGNATURE_CANDIDAATE.match(markers[::-1])
+    match = RE_SIGNATURE_CANDIDATE.match(markers[::-1])
     return candidate[-match.end('candidate'):] if match else []

talon/signature/learning/featurespace.py

@@ -1,6 +1,6 @@
 # -*- coding: utf-8 -*-
 
-""" The module provides functions for convertion of a message body/body lines
+""" The module provides functions for conversion of a message body/body lines
 into classifiers features space.
 
 The body and the message sender string are converted into unicode before
@@ -47,9 +47,9 @@ def apply_features(body, features):
     '''Applies features to message body lines.
 
     Returns list of lists. Each of the lists corresponds to the body line
-    and is constituted by the numbers of features occurances (0 or 1).
+    and is constituted by the numbers of features occurrences (0 or 1).
     E.g. if element j of list i equals 1 this means that
-    feature j occured in line i (counting from the last line of the body).
+    feature j occurred in line i (counting from the last line of the body).
     '''
     # collect all non empty lines
     lines = [line for line in body.splitlines() if line.strip()]
@@ -66,7 +66,7 @@ def build_pattern(body, features):
     '''Converts body into a pattern i.e. a point in the features space.
 
     Applies features to the body lines and sums up the results.
-    Elements of the pattern indicate how many times a certain feature occured
+    Elements of the pattern indicate how many times a certain feature occurred
     in the last lines of the body.
     '''
     line_patterns = apply_features(body, features)

talon/signature/learning/helpers.py

@@ -94,7 +94,7 @@ def binary_regex_match(prog):
 
 
 def flatten_list(list_to_flatten):
-    """Simple list comprehesion to flatten list.
+    """Simple list comprehension to flatten list.
 
     >>> flatten_list([[1, 2], [3, 4, 5]])
     [1, 2, 3, 4, 5]
@@ -155,7 +155,7 @@ def extract_names(sender):
 
 
 def categories_percent(s, categories):
-    '''Returns category characters persent.
+    '''Returns category characters percent.
 
     >>> categories_percent("qqq ggg hhh", ["Po"])
     0.0
@@ -177,7 +177,7 @@ def categories_percent(s, categories):
 
 
 def punctuation_percent(s):
-    '''Returns punctuation persent.
+    '''Returns punctuation percent.
 
     >>> punctuation_percent("qqq ggg hhh")
     0.0