Support import of various fixed-width formats.

This supports reading: SPSv1, SPSv2.1, P190 and custom
fixed-width formats. Supports skipping lines by startswith()
matching or by complete match (e.g., "EOF").

Closes #300 (SPS v1)
Closes #301 (SPS v2.1)
Closes #302 (P1/90)

bin/fwr.py

@@ -0,0 +1,108 @@
+#!/usr/bin/python3
+
+"""
+Fixed width record importing functions.
+"""
+
+import builtins
+
+transform = {
+	"int": lambda v: builtins.int(float(v)),
+	"float": float,
+	"string": str,
+	"bool": bool
+}
+
+def parse_line (line, fields):
+	data = dict()
+
+	for key in fields:
+		spec = fields[key]
+		transformer = transform[spec["type"]]
+		pos_from = spec["offset"]
+		pos_to = pos_from + spec["length"]
+		text = line[pos_from:pos_to]
+		value = transformer(text)
+		if "enum" in spec:
+			if "default" in spec:
+				value = spec["default"]
+			for enum_key in spec["enum"]:
+				if enum_key == value:
+					enum_value = transformer(spec["enum"][enum_key])
+					value = enum_value
+					break
+
+		data[key] = value
+
+	return data
+
+
+specfields = {
+	"sps1": {
+		"line_name": { "offset": 1, "length": 16, "type": "int" },
+		"point_number": { "offset": 17, "length": 8, "type": "int" },
+		"easting": { "offset": 46, "length": 9, "type": "float" },
+		"northing": { "offset": 55, "length": 10, "type": "float" }
+	},
+	"sps21": {
+		"line_name": { "offset": 1, "length": 7, "type": "int" },
+		"point_number": { "offset": 11, "length": 7, "type": "int" },
+		"easting": { "offset": 46, "length": 9, "type": "float" },
+		"northing": { "offset": 55, "length": 10, "type": "float" }
+	},
+	"p190": {
+		"line_name": { "offset": 1, "length": 12, "type": "int" },
+		"point_number": { "offset": 19, "length": 6, "type": "int" },
+		"easting": { "offset": 46, "length": 9, "type": "float" },
+		"northing": { "offset": 55, "length": 9, "type": "float" }
+	},
+}
+
+def from_file(path, spec):
+
+	# If spec.fields is not present, deduce it from spec.type ("sps1", "sps21", "p190", etc.)
+	if "fields" in spec:
+		fields = spec["fields"]
+	elif "type" in spec and spec["type"] in specfields:
+		fields = specfields[spec["type"]]
+	else:
+		return "Neither 'type' nor 'fields' given. I don't know how to import this fixed-width dataset."
+
+	firstRow = spec.get("firstRow", 0)
+
+	skipStart = [] # Skip lines starting with any of these values
+	skipMatch = [] # Skip lines matching any of these values
+
+	if "type" in spec:
+		if spec["type"] == "sps1" or spec["type"] == "sps21" or spec["type"] == "p190":
+			skipStart = "H"
+			skipMatch = "EOF"
+
+	records = []
+	with open(path, "r", errors="ignore") as fd:
+		row = 0
+		line = fd.readline()
+
+		while line:
+			skip = False
+
+			if row < firstRow:
+				skip = True
+
+			if not skip:
+				for v in skipStart:
+					if line.startswith(v):
+						skip = True
+						break
+				for v in skipMatch:
+					if line == v:
+						skip = True
+						break
+
+			if not skip:
+				records.append(parse_line(line, fields))
+
+			row += 1
+			line = fd.readline()
+
+	return records

bin/legacy_fwr.py

@@ -1,68 +0,0 @@
-#!/usr/bin/python3
-
-"""
-Legacy fixed width record importing functions.
-"""
-
-import builtins
-
-def parse_fwr (string, widths, start=0):
-	"""Parse a fixed-width record.
-
-	string: the string to parse.
-	widths: a list of record widths. A negative width denotes a field to be skipped.
-	start: optional start index.
-
-	Returns a list of strings.
-	"""
-	results = []
-	current_index = start
-	for width in widths:
-		if width > 0:
-			results.append(string[current_index : current_index + width])
-			current_index += width
-		else:
-			current_index -= width
-
-	return results
-
-def int (v):
-	return builtins.int(float(v))
-
-def parse_line (string, spec):
-	"""Parse a line from an SPS file."""
-	names = spec["names"]
-	widths = spec["widths"]
-	normalisers = spec["normalisers"]
-	record = [ t[0](t[1]) for t in zip(normalisers, parse_fwr(string, widths)) ]
-	return dict(zip(names, record))
-
-def from_file(path, spec = None):
-	if spec is None:
-		spec = {
-			"names": [ "line_name", "point_number", "easting", "northing" ],
-			"widths": [ -1, 10, 10, -25, 10, 10 ],
-			"normalisers": [ int, int, float, float ]
-		}
-	else:
-		normaliser_tokens = [ "int", "float", "str", "bool" ]
-		spec["normalisers"] = [ eval(t) for t in spec["types"] if t in normaliser_tokens ]
-
-	records = []
-	with open(path) as fd:
-		cnt = 0
-		line = fd.readline()
-		while line:
-			cnt = cnt+1
-
-			if line == "EOF":
-				break
-
-			record = parse_line(line, spec)
-			if record is not None:
-				records.append(record)
-
-			line = fd.readline()
-
-	del spec["normalisers"]
-	return records

bin/preplots.py

@@ -1,15 +1,33 @@
-import legacy_fwr
+import fwr
 
 """
 Preplot importing functions.
 """
 
 def from_file (file, realpath = None):
+	"""
+	Return a list of dicts, where each dict has the structure:
+		{
+			"line_name": <int>,
+			"points": [
+				{
+					"line_name": <int>,
+					"point_number": <int>,
+					"easting": <float>,
+					"northing": <float>
+				},
+				…
+			]
+		}
+	On error, return a string describing the error condition.
+	"""
+
 	filepath = realpath or file["path"]
-	if not "type" in file or file["type"] == "sps":
+	records = fwr.from_file(filepath, file)
-		records = legacy_fwr.from_file(filepath, file["format"] if "format" in file else None )
+
-	else:
+	if type(records) == str:
-		return "Not an SPS file"
+		# This is an error message
+		return records
 
 	lines = []
 	line_names = set([r["line_name"] for r in records])