fix: sub_array_range

Signed-off-by: 117503445 <[email protected]>
117503445 · Jun 27, 2023 · 2345c73 · 2345c73
1 parent a9de325
commit 2345c73
Show file tree

Hide file tree

Showing 3 changed files with 31 additions and 32 deletions.
diff --git a/config.yaml.example b/config.yaml.example
@@ -1,5 +1,5 @@
 path:
-  input: ./data
+  input: ./data/in
   output: /tmp/flow-pdf/out
 
 compare:

diff --git a/flow_pdf/worker/dump.py b/flow_pdf/worker/dump.py
@@ -80,7 +80,7 @@ def run_page(  # type: ignore[override]
             # for block in page_in.raw_dict["blocks"]:
             #     rects = []
             #     if block["type"] == 0:
-            #         for line in block["lines"]:
+            #         for line in block.lines:
             #             rects.append(line.bbox)
             #     add_annot(page, rects, "", "red")
             # add_annot(page, rects, "l", "red")
@@ -89,40 +89,39 @@ def run_page(  # type: ignore[override]
             # for block in page_in.raw_dict["blocks"]:
             #     rects = []
             #     if block["type"] == 0:
-            #         for line in block["lines"]:
-            #             for span in line["spans"]:
+            #         for line in block.lines:
+            #             for span in line.spans:
             #                 rects.append(span.bbox)
             #     add_annot(page, rects, "", "purple")
 
             # block common span
             # for block in page_in.raw_dict["blocks"]:
             #     rects = []
             #     if block["type"] == 0:
-            #         for line in block["lines"]:
-            #             for span in line["spans"]:
+            #         for line in block.lines:
+            #             for span in line.spans:
             #                 if is_common_span(span, doc_in.most_common_font, doc_in.common_size_range):
             #                     rects.append(span.bbox)
             #     add_annot(page, rects, "", "purple")
 
             # block not common span
-            # rects = []
-            # for c in page_in.big_blocks:
-            #     for block in c:
-            #         if block["type"] == 0:
-            #             for line in block["lines"]:
-            #                 for span in line["spans"]:
-            #                     if not is_common_span(span, doc_in.most_common_font, doc_in.common_size_range):
-            #                         rects.append(span.bbox)
-            # add_annot(page, rects, "", "purple")
+            rects = []
+            for blocks in page_in.big_blocks:
+                for block in blocks:
+                        for line in block.lines:
+                            for span in line.spans:
+                                if not is_common_span(span, doc_in.most_common_font, doc_in.common_size_range):
+                                    rects.append(span.bbox)
+            add_annot(page, rects, "", "purple")
 
             # new line
             # rects = []
             # for b in page_in.big_blocks:
-            #     for i in range(1, len(b["lines"])):
-            #         line = b["lines"][i]
+            #     for i in range(1, len(b.lines)):
+            #         line = b.lines[i]
             #         delta = line.bbox[0] - b.bbox[0]
             #         if delta > 1:
-            #             last_line = b["lines"][i - 1]
+            #             last_line = b.lines[i - 1]
             #             if last_line.bbox[0] - b.bbox[0] < 1:
             #                 rects.append(line.bbox)
             # add_annot(page, rects, "new-line", "pink")

diff --git a/flow_pdf/worker/font_counter.py b/flow_pdf/worker/font_counter.py
@@ -118,21 +118,21 @@ def sub_array_range(arr: list[float], sub_arr_range: int) -> Range:
 
                 arr.sort()
 
-                start = end = max_start = max_end = 0
-                max_count = count = 1
-
-                for i in range(1, len(arr)):
-                    if arr[i] - arr[start] < sub_arr_range:
-                        end = i
-                        count += 1
+                max_start = 0
+                max_end = 0
+                max_len = 1
+
+                start = 0
+                end = 0
+                while end < len(arr):
+                    if arr[end] - arr[start] > sub_arr_range:
+                        start += 1
                     else:
-                        start = end = i
-                        count = 1
-
-                    if count > max_count:
-                        max_count = count
-                        max_start = start
-                        max_end = end
+                        if end - start + 1 > max_len:
+                            max_len = end - start + 1
+                            max_start = start
+                            max_end = end
+                        end += 1
 
                 return Range(arr[max_start], arr[max_end])