Initial commit

.gitignore

+*.py[cod]
+__pycache__/
+*.so
+*~

asccol.py

+"""Parse AMRC's ASCII columnar data."""
+
+from collections import namedtuple
+
+Column = namedtuple('Column', ('name', 'type', 'width'))
+
+
+class DataSpec:
+    """Details about how to parse an aligned data file.
+
+    Arguments:
+        cols (Columns): columns' name, type, and width
+        leading (int): number of leading lines to ignore
+    """
+
+    def __init__(self, cols, leading=0):
+        self.cols = cols
+        self.leading = leading
+        self.namedtuple = namedtuple('row', (col.name for col in cols))
+
+
+class Columns(tuple):
+    """Details about how to parse aligned data."""
+
+    def __new__(cls, *data):
+        columns = map(Column._make, data)
+        return tuple.__new__(cls, columns)
+
+    def __repr__(self):
+        contents = ', '.join(map(repr, self))
+        return f'{type(self).__name__}({contents})'
+
+
+def parse_data(stream, dataspec):
+    """Read lines from `stream` and yield `namedtuple` data."""
+    for line_num, line in enumerate(stream, start=1):
+        if line_num <= dataspec.leading:
+            continue
+        row = parse_line(line, dataspec, line_num=line_num)
+        yield dataspec.namedtuple._make(row)
+
+
+def parse_line(line, dataspec, line_num=None):
+    """Yield data from the line."""
+    start = end = 0
+    for col_num, col in enumerate(dataspec.cols, start=1):
+        end += col.width
+        cell = line[start:end].strip()
+        if cell:
+            try:
+                yield col.type(cell)
+            except ValueError as err:
+                where = f'data column {col_num}'
+                if line_num is not None:
+                    where = f'line {line_num}, {where}'
+                message = f'Unexpected data format at {where}: {cell!r}'
+                raise ValueError(message) from err
+        else:
+            yield None
+
+        start = end