Draft class already does a good parsing and cleaning of the text so we can retrieve the abstract in a KISS way. Fixes ietf-tools#587

 - Legacy-Id: 2841

Author: emiliosanchez

ietf/utils/draft.py

@@ -277,73 +277,27 @@ def get_creation_date(self):
 
     # ------------------------------------------------------------------
     def get_abstract(self):
-        """Extract abstract from draft text.
-        based on: http://wiki.tools.ietf.org/tools/ietfdb/browser/branch/legacy/idst/validate.cgi
-        """
         if self._abstract:
             return self._abstract
-        draft_text = "\n".join(self.lines[:len(self.lines)/2])
-
-        abstract_patterns = [
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\s*Status of this Memo\s{0,2}\n)',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\s*Editorial Note\s{0,2}\n)',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\s*Sub-IP ID Summary\s{0,2}\n)',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\s*Table of Contents?\s{0,2}\n)',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\s*Conventions used in this document\s{0,2}\n)',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\s*Introduction\s{0,2}\n)',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\s*Terminology\s{0,2}\n)',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\s*Definitions\s\s{0,2}\n)',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\s*Background( and Intended Usage)?\s{0,2}\n)',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\s*Conventions\s{0,2}\n)',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\s*Requirements\s{0,2}\n)',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\s*motivation\s{0,2}\n)',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\s*translation\s{0,2}\n)',
-            r'((Abstract|Summary):?\n+)([\s|\S]+?)(\n\s*Contents?)',
-            r'((Abstract|Summary):?\n+)([\s|\S]+?)(\n\s*0. meta information on this)',
-            r'((Abstract|Summary):?\n+)([\s|\S]+?)(\n\s*0. introduction)',
-            r'((Abstract|Summary):?\s*\n+)(( {3,}.+\n+)+)',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+?)((\d\.?)?\n{4,})',
-            r'((Abstract|Summary):?\s*\n+)([\s|\S]+)((\d\.?)[^\n]{10,50}\n{0,2})',
-        ]
-
-        abstract = None
-        for ap in abstract_patterns:
-            match = re.search(ap, draft_text, re.IGNORECASE)
-            if match:
-                abstract = match.group(3)
-                break
-
-        # Remove possible missed sections
-        match = re.search(r'([\s|\S]+?)(\n+\s{0,10}(\d\.?)\s{0,10}[^\n]{10,50}\s{0,2}\n)', abstract or '')
-        if match:
-            abstract = match.group(1)
-
-        if abstract:
-            cleanup_patterns = [
-                (r"\n{3,}", "\n\n"), (r"\f[\n ]*[^\n]*\n", ""),
-                (r"(?s)(\d{1,3}\.\s*)?(Conventions used in this document|Requirements language).*", ""),
-                (r"(?sm)[\n ]*The key words [\'\"]MUST[\'\"], [\'\"]MUST NOT[\'\"],\n.*$", ""),
-                (r"(?s)\nStatus of [tT]his Memo\n\s{1,2}\n.*$", ""),
-                (r"(?s)(\d{1,3}\.?\s*)?Table of Contents\s{0,2}.*$", ""),
-                (r"(?s)(\d{1,3}\.?\s*)?Status of this memo\s{1,2}\n.*$", ""),
-                (r"(?s)(\d{1,3}\.?\s*)?Editorial note.\n*$", ""),
-                (r"(?s)(\d{1,3}\.?\s*)?Terminology\s{1,2}\n.*$", ""),
-                (r"(?s)(\d{1,3}\.?\s*)?Specification of Requirements\s{1,2}\n.*$", ""),
-                (r"(?s)\s*\d\.?(Abstract|Introduction)?\.{15,}", ""),
-                (r"\n\s*\d\.?\s*[\s|\S]+?\n.*", ""),
-                (r"(?sm)(\(|\[)?to be (deleted|removed) (by|when).*", ""),
-                (r"(?s)\n.$", ""), (r"(?m)^[^\n]+\.{10,}\d{1,3}", "")
-            ]
-            for cp, sub in cleanup_patterns:
-                e = re.compile(cp, re.IGNORECASE)
-                abstract = e.sub(sub, abstract)
-
-            # wrap long lines without messing up formatting
-            while re.match("([^\n]{72,}?) +", abstract):
-                abstract = re.sub("([^\n]{72,}?) +([^\n ]*)(\n|$)", "\\1\n\\2 ", abstract)
-
-            return abstract.strip()
-        return None
+        abstract_re = re.compile('^\s*abstract', re.I)
+        identation_re = re.compile('^(\s)*')
+
+        begin = False
+        identation = 0
+        abstract = []
+        for line in self.lines:
+            if abstract_re.match(line):
+                begin=True
+                continue
+            if begin:
+                if line and not line.startswith('  '):
+                    break
+                new_identation = len(identation_re.match(line).group(0))
+                if new_identation < identation:
+                    break
+                identation = new_identation
+                abstract.append(line)
+        return '\n'.join(abstract)
 
 
     # ------------------------------------------------------------------