Refactored table parsing.

main.py

@@ -1,6 +1,42 @@
+from typing import Optional, Dict
 import re
 import sys
 
+def parse_table_entries(fields, pattern, lines):
+    """Parse aligned table entries"""
+    prev: Optional[Dict] = None
+    entries = []
+    def flush():
+        nonlocal prev
+        if prev is not None:
+            entries.append(prev['data'])
+            prev = None
+    for line in lines:
+        if len(line.strip()) == 0:
+            flush()
+            continue
+
+        match = pattern.search(line.strip())
+        if match is not None:
+            flush()
+            starts = [match.span(v)[0] for v in fields] + [len(line)]
+            prev = {
+                'data': {v: match.group(v) for v in fields},
+                'starts': starts
+            }
+        elif prev is not None:
+            for i, field in enumerate(fields):
+                v = line[prev['starts'][i]:prev['starts'][i + 1]].strip()
+                if len(v) == 0:
+                    continue
+                if prev['data'][field][-1] not in ['-', '_']:
+                    v = ' ' + v
+                prev['data'][field] += v
+
+    flush()
+    return entries
+
+
 def parse_registers(lines):
     """Parse register entry from `pdftotext -layout output` from TRM table"""
     fields = ['id', 'address', 'size', 'type', 'reset_value', 'description']
@@ -12,37 +48,7 @@ def parse_registers(lines):
                                r'(?P<type>\w+)\s+'
                                r'(?P<reset_value>0x[0-9A-F]+)\s+'
                                r'(?P<description>.+)$')
-    prev = None
+    return parse_table_entries(fields, ENTRY_PATTERN, lines)
-    registers = []
-    for line in lines:
-        if len(line.strip()) == 0:
-            if prev is not None:
-                registers.append(prev['reg'])
-            prev = None
-            continue
-
-        match = ENTRY_PATTERN.search(line.strip())
-        if match is not None:
-            if prev is not None:
-                registers.append(prev['reg'])
-            starts = [match.span(v)[0] for v in fields] + [len(line)]
-            prev = {
-                'reg': {v: match.group(v) for v in fields},
-                'starts': starts
-            }
-        elif prev is not None:
-            for i, field in enumerate(fields):
-                v = line[prev['starts'][i]:prev['starts'][i + 1]].strip()
-                if len(v) == 0:
-                    continue
-                if prev['reg'][field][-1] not in ['-', '_']:
-                    v = ' ' + v
-                prev['reg'][field] += v
-
-    if prev is not None:
-        registers.append(prev['reg'])
-
-    return registers
 
 def emit_rust(base_addr, ending_addr, registers):
     current_addr = base_addr