Merge branch 'refactor_align' of horel.org:thibaut/ocr-layer-curation into refactor_align

1 files changed, 11 insertions, 0 deletions

diff --git a/compare.py b/compare.py
index b6a4c8a..daec25c 100644
--- a/compare.py
+++ b/compare.py
@@ -14,8 +14,19 @@ l1, c1 = ocrpage['words'], ocrpage["coords"]
 l2 = get_page(wikibook, n)
 print len(l2.split())
 l3 = su.simplify(l2)
+
+def del_cost1(w, pos):
+    return 50
+
+def del_cost2(w, pos):
+    return 1+3*len([c for c in w if c.isalnum()])
+bactrack1 = 8
+backtrack2 = 5
+
 C = su.align(l3.split(), l1, c1)
 pdb.set_trace()
+su.print_alignment(l2.split(), l1, c1, C[1])
+
 sexp = page_sexp(wikibook, n)
 su.alignment_to_sexp(C[1], sexp, l2.split())
 su.print_alignment(l2.split(), l1, c1, C[1])