First work towards SPMD support.

4 files changed, 26 insertions, 143 deletions

diff --git a/README.markdown b/README.markdown
index 1fea850..a561dcb 100644
--- a/README.markdown
+++ b/README.markdown
@@ -35,131 +35,4 @@ Thanks for the inspiration goes to:
 Input format
 ============
 
-The format of the text input is very much like how you would write a
-screenplay on and old typewriter, only that you don't have to worry about tab
-stops and line breaks.
-
-Here's an example:
-
-    EXT. CASTLE WALLS - DAY
-
-    Mist. Several seconds of it swirling about.
-    Silence possibly, atmospheric music. SUPERIMPOSE "England AD 787".
-    After a few more seconds we hear hoofbeats in
-    the distance.
-    They come slowly closer. Then out of the mist comes KING ARTHUR
-    followed by a SERVANT who is banging two half coconuts
-    together. ARTHUR raises his hand.
-
-    ARTHUR
-    Whoa there!
-
-
-    SERVANT makes noises of horses halting, with a flourish. ARTHUR
-    peers through the mist. CUT TO shot from over his shoulder:
-    castle (e.g. Bodium) rising out of the mist. On the castle
-    battlements a SOLDIER is dimly seen. He peers down.
-
-     SOLDIER
-     Halt! Who goes there?
-
-     ARTHUR
-     It is I, Arthur, son of Uther Pendragon,
-     from the castle
-    of Camelot.
-      King of all Britons, defeator of the Saxons,
-    sovereign of all England!
-
-Note the free (that is, pretty ugly) format of this input.
-After Screenplain has digested it, it will appear correctly formatted like this:
-
-    EXT. CASTLE WALLS - DAY
-
-    Mist. Several seconds of it swirling about. Silence possibly,
-    atmospheric music. SUPERIMPOSE "England AD 787". After a few more
-    seconds we hear hoofbeats in the distance. They come slowly closer.
-    Then out of the mist comes KING ARTHUR followed by a SERVANT who is
-    banging two half coconuts together. ARTHUR raises his hand.
-
-                          ARTHUR
-              Whoa there!
-
-    SERVANT makes noises of horses halting, with a flourish. ARTHUR
-    peers through the mist. CUT TO shot from over his shoulder: castle
-    (e.g. Bodium) rising out of the mist. On the castle battlements a
-    SOLDIER is dimly seen. He peers down.
-
-                           SOLDIER
-              Halt! Who goes there?
-
-                           ARTHUR
-              It is I, Arthur, son of Uther
-              Pendragon, from the castle of
-              Camelot. King of all Britons,
-              defeator of the Saxons, sovereign
-              of all England!
-
-As you can see, you can write your screenplay without keeping formatting rules in your head.
-As Screenplain is just a piece of software, and not a mind-reading robot, it does have rules,
-but they are designed to be intuitive and shouldn't distract from the writing.
-
-Basically, what Screenplain does with your plain text file is to split it into paragraphs.
-A paragraph is simply one or more lines of text. Paragraphs are separated by one or
-more empty lines. The example above contains six paragraphs.
-
-After splitting the text into paragraphs, Screenplain decides what type each paragraph has.
-It can be one of:
-
-  * Slug line
-  * Dialogue
-  * Action
-
-The paragraph type decides how the paragraph will look in the output. See
-below for how Screenplain figures out which one of these types each paragraph has.
-
-Slug lines
-----------
-
-If a paragraph contains *one single line* written in *capital letters*, Screenplain
-assumes it is a slug line.
-
-A slug line is a line that starts a scene. An example of a slug line is `EXT.
-EIFFEL TOWER - DAY`, but Screenplain does not care about the format. Screenplain
-assumes `SOMEWHERE IN SPACE` is a slug line too.
-
-Dialogue
---------
-
-If a paragraph consists of two or more lines, and the first line is written
-only in *capital letters*, Screenplain assumes it is a piece of dialogue. The
-first line should contain the name of the speaking character. The rest is
-whatever that character says. Simple isn't it? Well, there is an exception: If
-a line starts with an opening parenthesis, "(", it starts a parenthetical
-block, which is a screenwriter's way of telling the actor how the line is to
-be delivered (something that pisses directors off). Parentheticals are
-formatted differently from the spoken lines.
-
-Action
-------
-
-Any paragraph that is not one of the other types contains descriptions about
-what happens in the scene.
-
-TO DO
-=====
-
-  * Specify options on the command line. Duh.
-
-  * Installation package.
-
-  * Margins in PDF output.
-
-  * Support for transitions. A `CUT TO:` line is interpreted as a slug line.
-    Fixing this should be simple: any single-line paragraph that starts with one
-    or more spaces and does not contain lower-case characters should be a
-    transition. Any pitfalls with that? What about left-aligned transitions like
-    `FADE IN:`?
-
-  * Good support for national characters. It works on my system, but I'm not so
-    sure if it will work for everyone. Should we just support UTF-8, or allow
-    other encodings?
+See http://prolost.com/storage/downloads/spmd/SPMD_proposal.html
diff --git a/screenplain/export/text.py b/screenplain/export/text.py
index 49a1f23..e135646 100644
--- a/screenplain/export/text.py
+++ b/screenplain/export/text.py
@@ -1,11 +1,11 @@
 import sys
 from screenplain.parse import parse, get_pages
 
-def to_text(input):
-    out = sys.stdout
+def to_text(input, out):
     paragraphs = parse(input)
-    for page_no, page in enumerate(get_pages(paragraphs), 1):
-        if page_no != 1:
+    for page_no, page in enumerate(get_pages(paragraphs)):
+        # page_no is 0-based
+        if page_no != 0:
             out.write('\f')
         for line in page:
             out.write(line)
diff --git a/screenplain/main.py b/screenplain/main.py
index 948914c..bbe6450 100644
--- a/screenplain/main.py
+++ b/screenplain/main.py
@@ -7,9 +7,9 @@ import fileinput
 import sys
 
 if __name__ == '__main__':
-    if False:
+    if True:
         from screenplain.export.text import to_text
-        to_text(fileinput.input())
+        to_text(fileinput.input(), sys.stdout)
     else:
         from screenplain.export.pdf import to_pdf
         to_pdf(fileinput.input())
diff --git a/screenplain/parse.py b/screenplain/parse.py
index c4cadb5..a872f7e 100644
--- a/screenplain/parse.py
+++ b/screenplain/parse.py
@@ -1,10 +1,14 @@
 import itertools
 import textwrap
+import re
 
 # Numbers from http://www.emacswiki.org/emacs/ScreenPlay
 # According to http://johnaugust.com/2004/how-many-lines-per-page
 lines_per_page = 56
 
+slug_re = re.compile('^(EXT\.|INT\.|EXT/INT\.|I/E) ', re.IGNORECASE)
+
+
 class _ParentheticalFlipFlop(object):
     def __init__(self):
         self.inside = False
@@ -85,24 +89,30 @@ class Action(object):
             yield self.indent + line
 
 def is_blank(string):
-    return string == '' or string.isspace()
+    return string == '' or string.isspace() and string != '  '
 
-def create_paragraph(line_list):
-    if line_list[0].isupper():
-        if len(line_list) == 1:
-            return Slug(line_list)
-        else:
-            return Dialog(line_list)
+def is_slug(blanks_before, string):
+    return blanks_before >= 2 or slug_re.match(string)
+
+def create_paragraph(blanks_before, line_list):
+    if is_slug(blanks_before, line_list[0]):
+        return Slug(line_list)
+    if len(line_list) > 1 and line_list[0].isupper():
+        return Dialog(line_list)
     else:
         return Action(line_list)
 
 def parse(source):
     """Reads raw text input and generates paragraph objects."""
+    blank_count = 0
     for blank, lines in itertools.groupby(
         (line.rstrip() for line in source), is_blank
     ):
-        if not blank:
-            yield create_paragraph(list(lines))
+        if blank:
+            blank_count += 1
+        else:
+            yield create_paragraph(blank_count, list(lines))
+            blank_count = 0
 
 def get_pages(paragraphs):
     """Generates one list of lines per page."""