Memo preamble parsing

jubeatools/formats/__init__.py

@@ -16,6 +16,8 @@ from .memon import (
 
 ALIASES = {
     "memon": "memon:v0.2.0",
+    "ichi-retsu": "mono-column",
+    "memo": "memo2",
 }
 
 # Loaders deserialize a folder or a file to a Song object

jubeatools/formats/memo/__init__.py

@@ -0,0 +1,10 @@
+"""
+memo is a vague term refering to several legacy formats.
+They were originally derived from the (somewhat) human-readable format choosen
+by websites storing official jubeat charts in text form as a memory aid.
+
+The machine-readable variants are partially documented (in japanese)
+on these pages :
+- http://yosh52.web.fc2.com/jubeat/fumenformat.html
+- http://yosh52.web.fc2.com/jubeat/holdmarker.html
+"""

jubeatools/formats/memo/preamble.py

@@ -0,0 +1,79 @@
+"""
+Useful things to parse the preamble of analyser-like formats
+"""
+from decimal import Decimal
+from typing import List, Tuple, Union
+
+from parsimonious import Grammar, NodeVisitor
+from parsimonious.expressions import Node
+
+preamble_line_grammar = Grammar(
+    r"""
+    raw_line        = line comment?
+    line            = command? ws
+    command         = hash_command / simple_command
+    hash_command    = "#" key "=" value
+    key             = ~"[a-z]+"
+    value           = value_in_quotes / number
+    value_in_quotes = "\"" quoted_value "\""
+    quoted_value    = ~"[^\"]+"
+    number          = ~"\d+"
+    simple_command  = letter "=" value
+    letter          = ~"\w"
+    ws              = ~"[\t ]*"
+    comment         = ~"//.*"
+    """
+)
+
+
+class PreambleLineVisitor(NodeVisitor):
+
+    """Returns a (key, value) tuple or None if the line contains no useful
+    information for the parser (a comment or an empty line)"""
+
+    def _as_text(self, node, visited_children):
+        return node.text
+
+    def visit_raw_line(self, node, visited_children):
+        value, _ = visited_children
+        return value
+
+    def visit_line(self, node, visited_children):
+        command, _ = visited_children
+        value = list(command)
+        return value[0] if value else None
+
+    visit_command = NodeVisitor.lift_child
+
+    def visit_hash_command(self, node, visited_children):
+        _, key, _, value = visited_children
+        return (key, value)
+
+    visit_key = _as_text
+
+    visit_value = NodeVisitor.lift_child
+
+    def visit_value_in_quotes(self, node, visited_children):
+        _, value, _ = visited_children
+        return value
+
+    visit_quoted_value = _as_text
+
+    def visit_number(self, node, visited_children):
+        return Decimal(node.text)
+
+    def visit_simple_command(self, node, visited_children):
+        letter, _, value = visited_children
+        return (letter, value)
+
+    visit_letter = _as_text
+
+    def generic_visit(self, node, visited_children):
+        return visited_children or node
+
+
+_preamble_line_visitor = PreambleLineVisitor()
+
+
+def load_preamble_line(line: str) -> Union[Tuple[str, Union[str, Decimal]], None]:
+    _preamble_line_visitor.visit(preamble_line_grammar.parse(line))

poetry.lock

@@ -108,6 +108,17 @@ version = "20.4"
 pyparsing = ">=2.0.2"
 six = "*"
 
+[[package]]
+category = "main"
+description = "(Soon to be) the fastest pure-Python PEG parser I could muster"
+name = "parsimonious"
+optional = false
+python-versions = "*"
+version = "0.8.1"
+
+[package.dependencies]
+six = ">=1.9.0"
+
 [[package]]
 category = "main"
 description = "A module wrapper for os.path"
@@ -205,7 +216,7 @@ python-versions = ">=2.5, !=3.0.*, !=3.1.*, !=3.2.*"
 version = "3.17.0"
 
 [[package]]
-category = "dev"
+category = "main"
 description = "Python 2 and 3 compatibility utilities"
 name = "six"
 optional = false
@@ -237,7 +248,7 @@ python-versions = "*"
 version = "0.1.9"
 
 [metadata]
-content-hash = "b55c6b2244d11c0356dc49e32615244f2b964ac0b903d8613a5e0e453557643f"
+content-hash = "a07afd18095d3d59ba9a855d90c797920c63fd004b7e1cb134a5ec981e7e16b8"
 python-versions = "^3.8"
 
 [metadata.files]
@@ -296,6 +307,9 @@ packaging = [
     {file = "packaging-20.4-py2.py3-none-any.whl", hash = "sha256:998416ba6962ae7fbd6596850b80e17859a5753ba17c32284f67bfff33784181"},
     {file = "packaging-20.4.tar.gz", hash = "sha256:4357f74f47b9c12db93624a82154e9b120fa8293699949152b22065d556079f8"},
 ]
+parsimonious = [
+    {file = "parsimonious-0.8.1.tar.gz", hash = "sha256:3add338892d580e0cb3b1a39e4a1b427ff9f687858fdd61097053742391a9f6b"},
+]
 path = [
     {file = "path-14.0.1-py3-none-any.whl", hash = "sha256:ed305ab0629859795863afdcd2ff8950af9c1d703527cb8b8aec3b6a425a80ba"},
     {file = "path-14.0.1.tar.gz", hash = "sha256:aedf5be64bcfe34dcc17b2bcf379e86f8f5e4d06588b3697e1bba4a1563794b4"},

pyproject.toml

@@ -12,6 +12,7 @@ click = "^7.1.2"
 path = "^14.0.1"
 simplejson = "^3.17.0"
 marshmallow = "^3.6.0"
+parsimonious = "^0.8.1"
 
 [tool.poetry.dev-dependencies]
 pytest = "^5.2"