aboutsummaryrefslogtreecommitdiffstats
path: root/web/utils.py
diff options
context:
space:
mode:
Diffstat (limited to 'web/utils.py')
-rw-r--r--web/utils.py21
1 files changed, 15 insertions, 6 deletions
diff --git a/web/utils.py b/web/utils.py
index bb9a4fe..8522841 100644
--- a/web/utils.py
+++ b/web/utils.py
@@ -1,12 +1,21 @@
-from parsedjvutext import parse_book
+import djvu_utils as du
import sys
-
+import string_utils as su
+from wikisource import get_page
def gen_html(book, page_number):
- d = parse_book(book, page=int(page_number), html=True)
- if d[0]:
- words, coords = zip(*d[0])
- return (list(enumerate(coords)), list(enumerate(words)))
+ doc = du.get_document("../" + book)
+ page = doc.pages[int(page_number)-1]
+ d = du.parse_page(page)
+ corrected_text = get_page(book, int(page_number))
+ corrected_words = su.simplify(corrected_text).split()
+ if d:
+ words, coords = zip(*d)
+ C = su.align(corrected_words, list(words), list(coords))
+ r = su.alignment_to_sexp(corrected_text.split(), words, coords, C[1])
+ corrected_words, coords = zip(*r)
+ coords_html = du.convert_to_htmlcoord(coords, page.size[1])
+ return (list(enumerate(coords_html)), list(enumerate(corrected_words)))
if __name__ == "__main__":
gen_html(*sys.argv[1:3])