qurator-spk/eynollah
1
0
Fork 0
mirror of https://github.com/qurator-spk/eynollah.git synced 2025-10-06 22:50:14 +02:00
Code Issues Projects Releases Wiki Activity

get textlines sorted in textregion - verticals

Browse source
This commit is contained in:
b-vr103 2025-09-24 03:43:36 +02:00 • committed by kba
parent 6904a98182
commit 6d8641a518
1 changed files with 51 additions and 23 deletions
Download patch file Download diff file Expand all files Collapse all files

38
src/eynollah/eynollah.py
View file

@ -1759,14 +1759,19 @@ class Eynollah:
if N==0: if N==0:
return [] return []
diff_matrix = np.abs(np.subtract.outer(cy_textline, cy_textline)) diff_cy = np.abs( np.diff(sorted(cy_textline)) )
diff_cx = np.abs(np.diff(sorted(cx_textline)) )
non_zero_diffs = diff_matrix[diff_matrix > 0]
if len(non_zero_diffs) == 0: if len(diff_cy)>0:
mean_y_diff = 0 mean_y_diff = np.mean(diff_cy)
mean_x_diff = np.mean(diff_cx)
else: else:
mean_y_diff = np.mean(non_zero_diffs) mean_y_diff = 0
mean_x_diff = 0
if np.int(mean_y_diff) >= np.int(mean_x_diff):
row_threshold = mean_y_diff / 2 if mean_y_diff > 0 else 10 row_threshold = mean_y_diff / 2 if mean_y_diff > 0 else 10
indices_sorted_by_y = sorted(range(N), key=lambda i: cy_textline[i]) indices_sorted_by_y = sorted(range(N), key=lambda i: cy_textline[i])
@ -1789,6 +1794,29 @@ class Eynollah:
for idx in row_sorted: for idx in row_sorted:
sorted_textlines.append(textlines_textregion[idx]) sorted_textlines.append(textlines_textregion[idx])
else:
row_threshold = mean_x_diff / 2 if mean_x_diff > 0 else 10
indices_sorted_by_x = sorted(range(N), key=lambda i: cx_textline[i])
rows = []
current_row = [indices_sorted_by_x[0]]
for i in range(1, N):
current_idy = indices_sorted_by_x[i]
prev_idy = current_row[0]
if abs(cx_textline[current_idy] - cx_textline[prev_idy] ) <= row_threshold:
current_row.append(current_idy)
else:
rows.append(current_row)
current_row = [current_idy]
rows.append(current_row)
sorted_textlines = []
for row in rows:
row_sorted = sorted(row , key=lambda i: cy_textline[i])
for idy in row_sorted:
sorted_textlines.append(textlines_textregion[idy])
return sorted_textlines return sorted_textlines