From 10ce0e07dd33f8bf2a1354ee45d40207327318f5 Mon Sep 17 00:00:00 2001
From: Henrik Levkowetz <henrik@levkowetz.com>
Date: Sun, 10 Jun 2007 11:27:02 +0000
Subject: [PATCH] 'soup2text' is a html-to-text converter which uses the
 BeautifulSoup.py module.  It converts html to plain paragraph-filled readable
 text.  - Legacy-Id: 277

---
 ietf/utils/__init__.py  |  1 +
 ietf/utils/soup2text.py | 83 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 84 insertions(+)
 create mode 100755 ietf/utils/soup2text.py

diff --git a/ietf/utils/__init__.py b/ietf/utils/__init__.py
index b01f5b7df..69fd3bb07 100644
--- a/ietf/utils/__init__.py
+++ b/ietf/utils/__init__.py
@@ -2,6 +2,7 @@ from listop import orl, flattenl
 from log import log
 from cache_foreign_key import FKAsOneToOne
 from templated_form import makeTemplatedForm
+from soup2text import TextSoup, soup2text
 
 makeFormattingForm = makeTemplatedForm
 
diff --git a/ietf/utils/soup2text.py b/ietf/utils/soup2text.py
new file mode 100755
index 000000000..f0529d43b
--- /dev/null
+++ b/ietf/utils/soup2text.py
@@ -0,0 +1,83 @@
+#!/usr/bin/env python
+
+import re
+import textwrap
+from BeautifulSoup import Tag, BeautifulSoup, NavigableString
+
+block_tags = ["[document]", "html", "body", "div", "blockquote", "table", "tr", "p", "pre", "h1", "h2", "h3", "h4", "h5", "h6", ]
+ignore_tags = ["head", "script", "style"]
+pre_tags = ["pre"]
+entities = [("&lt;", "<"),   ("&gt;", ">"),
+            ("&quot;", '"'), ("&apos;", "'"),
+            ("&nbsp;", " "),
+            ("&amp;", "&"), ]
+
+def para(words, pre):
+    text = " ".join(words)
+    for entity, char in entities:
+        text = text.replace(entity, char)
+    if not pre:
+        text = re.sub("[\r\n\t ]+", " ", text)
+        text = textwrap.fill(text)
+    return text
+
+def render(node, encoding='latin-1', pre=False):
+    blocks = []
+    words = []
+    node.pre = pre or node.name in pre_tags
+    node.is_block = node.name in block_tags
+    for child in node:
+        if isinstance(child, NavigableString):
+            str = child.__str__(encoding)
+            if str and not node.pre:
+                str = str.strip()
+            if str and not str.startswith("<!") and not str.startswith("<?"):
+                words.append(str)
+        elif isinstance(child, Tag):
+            if child.name in ignore_tags:
+                pass
+            else:
+                child = render(child, encoding, node.pre)
+                if child.text:
+                    if child.is_block:
+                        if words :
+                            blocks.append(para(words, node.pre)+"\n")
+                            words = []
+                        blocks.append(child.text+"\n\n")
+                        node.is_block = True
+                    else:
+                        words.append(child.text)
+        else:
+            raise ValueException("Unexpected node type: '%s'" % child)
+    if words:
+        blocks.append(para(words, node.pre))
+
+    node.text = ''.join(blocks)
+    return node
+
+class TextSoup(BeautifulSoup):
+
+    def __str__(self, encoding='latin-1',
+                prettyPrint=False, indentLevel=0):
+        node = render(self, encoding)
+        str = node.text
+        str = re.sub("[ \t]+", " ", str)
+        str = re.sub("\n\n+", "\n\n", str)
+        return str
+
+def soup2text(html):
+    soup = TextSoup(html)
+    return str(soup)
+
+if __name__ == "__main__":
+    import sys
+    import urllib2 as urllib
+    for arg in sys.argv[1:]:
+        if arg[:6] in ["http:/", "https:", "ftp://"]:
+            file = urllib.urlopen(arg)
+        else:
+            file = open(arg)
+        html = file.read()
+        file.close
+        soup = TextSoup(html)
+        print str(soup)