feat: modular grammar

daidepp/parser/grammar.py

@@ -1,27 +1,350 @@
+"""
+For now this grammar assumes NO PDA games
+"""
+
+from __future__ import annotations
+
+from typing import Dict, List, Set, Tuple
+
 from parsimonious.grammar import Grammar
+from typing_extensions import Literal
+
+
+class DAIDEGrammar(Grammar):
+    def __init__(self, rules: str = "", **more_rules) -> None:
+        super().__init__(rules, **more_rules)
+        self._set_try_tokens()
+
+    def _set_try_tokens(self):
+        try_tokens = self.get("try_tokens")
+        try_tokens_strings = list(map(lambda x: x.literal, try_tokens.members))
+        self.try_tokens: List[str] = try_tokens_strings
+
+    @staticmethod
+    def from_level(level: DAIDE_LEVEL):
+        return create_daide_grammar(level)
+
+
+DAIDE_LEVEL = Literal[10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130]
+
+TRAIL_TOKEN = "---"  # any value starting with '---' is meant to be a continuation of that key, not a replacement
+
+# Peace and Alliances
+LEVEL_10: GRAMMAR_DICT = {
+    "pce": '"PCE" lpar power (ws power)* rpar',
+    "ccl": '"CCL" lpar press_message rpar',
+    "fct": '"FCT" lpar arrangement rpar',
+    "try": '"TRY" lpar try_tokens (ws try_tokens)* rpar',
+    "huh": '"HUH" lpar press_message rpar',
+    "prp": '"PRP" lpar arrangement rpar',
+    "aly_vss": '"ALY" lpar power (ws power)* rpar "VSS" lpar power (ws power)* rpar',
+    "slo": '"SLO" lpar power rpar',
+    "not": '"NOT" lpar arrangement rpar',
+    "nar": '"NAR" lpar arrangement rpar',
+    "drw": '"DRW" (lpar power (ws power)+ rpar)?',
+    "yes": '"YES" lpar press_message rpar',
+    "rej": '"REJ" lpar press_message rpar',
+    "bwx": '"BWX" lpar press_message rpar',
+    "fct": '("FCT" lpar arrangement rpar)',
+    "frm": '"FRM" lpar power rpar lpar power (ws power)* rpar lpar message rpar',
+    "power": '"AUS" / "ENG" / "FRA" / "GER" / "ITA" / "RUS" / "TUR"',
+    "lpar": '~"\s*\(\s*"',
+    "rpar": '~"\s*\)\s*"',
+    "ws": '~"\s+"',
+    "reply": "yes / rej / bwx / huh",
+    "message": "press_message / reply",
+    "press_message": "prp / ccl / fct / try / frm",
+    "arrangement": "pce / aly_vss / drw / slo / not / nar",
+    "try_tokens": '"PRP" / "PCE" / "ALY" / "VSS" / "DRW" / "SLO" / "NOT" / "YES" / "REJ" / "BWX" / "FCT"',
+}
+
+# Order Proposals
+LEVEL_20: GRAMMAR_DICT = {
+    "xdo": '"XDO" lpar order rpar',
+    "dmz": '"DMZ" lpar power (ws power)* rpar lpar province (ws province)* rpar',
+    "order": "hld / mto / sup / cvy / move_by_cvy / retreat / build",
+    "hld": 'lpar unit rpar "HLD"',
+    "mto": 'lpar unit rpar "MTO" ws province',
+    "sup": 'lpar unit rpar "SUP" lpar unit rpar ("MTO" ws prov_no_coast)?',
+    "cvy": 'lpar unit rpar "CVY" lpar unit rpar "CTO" ws province',
+    "move_by_cvy": 'lpar unit rpar "CTO" ws province ws "VIA" lpar prov_sea (ws prov_sea)* rpar',
+    "retreat": "rto / dsb",
+    "rto": 'lpar unit rpar "RTO" ws province',
+    "dsb": 'lpar unit rpar "DSB"',
+    "build": "bld / rem / wve",
+    "bld": 'lpar unit rpar "BLD"',
+    "rem": 'lpar unit rpar "REM"',
+    "wve": 'power ws "WVE"',
+    "unit_type": '"AMY" / "FLT"',
+    "unit": "power ws unit_type ws province",
+    "province": "prov_coast / prov_no_coast / prov_sea / (lpar prov_coast ws coast rpar)",
+    "coast": '"NCS" / "ECS" / "SCS" / "WCS"',
+    "prov_coast": '"ALB" / "ANK" / "APU" / "ARM" / "BEL" / "BER" / "BRE" / "BUL" / "CLY" / "CON" / "DEN" / "EDI" / "FIN" / "GAS" / "GRE" / "HOL" / "KIE" / "LON" / "LVN" / "LVP" / "MAR" / "NAF" / "NAP" / "NWY" / "PIC" / "PIE" / "POR" / "PRU" / "ROM" / "RUM" / "SEV" / "SMY" / "SPA" / "STP" / "SWE" / "SYR" / "TRI" / "TUN" / "TUS" / "VEN" / "YOR" / "WAL"',
+    "prov_no_coast": '"BOH" / "BUD" / "BUR" / "MOS" / "MUN" / "GAL" / "PAR" / "RUH" / "SER" / "SIL" / "TYR" / "UKR" / "VIE" / "WAR" ',
+    "prov_sea": '"ADR" / "AEG" / "BAL" / "BAR" / "BLA" / "BOT" / "EAS" / "ENG" / "HEL" / "ION" / "IRI" / "LYO" / "MAO" / "NAO" / "NTH" / "NWG" / "SKA" / "TYS" / "WES"',
+    "supply_center": '"ANK" / "BEL" / "BER" / "BRE" / "BUD" / "BUL" / "CON" / "DEN" / "EDI" / "GRE" / "HOL" / "KIE" / "LON" / "LVP" / "MAR" / "MOS" / "MUN" / "NAP" / "NWY" / "PAR" / "POR" / "ROM" / "RUM" / "SER" / "SEV" / "SMY" / "SPA" / "STP" / "SWE" / "TRI" / "TUN" / "VEN" / "VIE" / "WAR"',
+    "arrangement": f"{TRAIL_TOKEN}mto / xdo / dmz",
+    "try_tokens": f'{TRAIL_TOKEN}"XDO" / "DMZ"',
+}
+
+# Multipart Arrangements
+LEVEL_30: GRAMMAR_DICT = {
+    "and": '"AND" lpar sub_arrangement rpar (lpar sub_arrangement rpar)+',
+    "orr": '"ORR" lpar sub_arrangement rpar (lpar sub_arrangement rpar)+',
+    "sub_arrangement": "pce / aly_vss / drw / slo / not / nar / mto / xdo / dmz",
+    "arrangement": f"{TRAIL_TOKEN}and / orr",
+    "try_tokens": f'{TRAIL_TOKEN}"AND" / "ORR"',
+}
+
+# Sharing out Supply Centers
+LEVEL_40: GRAMMAR_DICT = {
+    "scd": '"SCD" (lpar power ws supply_center (ws supply_center)* rpar)+',
+    "occ": '"OCC" (lpar unit rpar)+',
+    "arrangement": f"{TRAIL_TOKEN}scd / orr",
+    "try_tokens": f'{TRAIL_TOKEN}"SCD" / "OCC"',
+}
+
+# Nested Multipart Arrangements
+LEVEL_50: GRAMMAR_DICT = {
+    "and": '"AND" lpar arrangement rpar (lpar arrangement rpar)+',
+    "orr": '"ORR" lpar arrangement rpar (lpar arrangement rpar)+',
+    "cho": '"CHO" lpar (~"\d+ \d+") rpar (lpar arrangement rpar)+',
+    "arrangement": f"{TRAIL_TOKEN}cho",
+    "try_tokens": f'{TRAIL_TOKEN}"CHO"',  # This isn't included in the originla daide spec but I think they just forgot it.
+}
+
+# Queries and Insistencies
+LEVEL_60: GRAMMAR_DICT = {
+    "ins": '"INS" lpar arrangement rpar',
+    "qry": '"QRY" lpar arrangement rpar',
+    "thk": '("THK" lpar qry rpar) / ("THK" lpar "NOT" lpar qry rpar rpar) / ("THK" lpar arrangement rpar)',
+    "idk": '"IDK" lpar qry rpar',
+    "sug": '"SUG" lpar arrangement rpar',
+    "fct": '("FCT" lpar qry rpar) / ("FCT" lpar "NOT" lpar qry rpar rpar) / ("FCT" lpar arrangement rpar)',
+    "reply": f"{TRAIL_TOKEN}fct / thk / idk",
+    "try_tokens": f'{TRAIL_TOKEN}"INS" / "QRY" / "IDK" / "IDK" / "SUG"',
+    "press_message": f"{TRAIL_TOKEN}thk / ins / qry / sug",
+}
+
+# Requests for Suggestion
+LEVEL_70: GRAMMAR_DICT = {
+    "wht": '"WHT" lpar unit rpar',
+    "how": '("HOW" lpar province rpar) / ("HOW" lpar power rpar)',
+    "try_tokens": f'{TRAIL_TOKEN}"HOW" / "WHT"',
+    "press_message": f"{TRAIL_TOKEN}wht / how",
+}
+
+# Accusations
+LEVEL_80: GRAMMAR_DICT = {
+    "exp": '"EXP" lpar turn rpar lpar message rpar',
+    "idk": '("IDK" lpar exp rpar) / ("IDK" lpar qry rpar)',
+    "sry": '"SRY" lpar exp rpar',
+    "turn": 'season ws ~"\d{4}"',
+    "season": '"SPR" / "SUM" / "FAL" / "AUT" / "WIN"',
+    "press_message": f"{TRAIL_TOKEN}exp",
+    "try_tokens": f'{TRAIL_TOKEN}"EXP" / "SRY"',
+}
+
+# Future Discussions
+LEVEL_90: GRAMMAR_DICT = {
+    "for": '("FOR" lpar turn rpar lpar arrangement rpar) / ("FOR" lpar (lpar turn rpar lpar turn rpar) rpar lpar arrangement rpar)',
+    "arrangement": f"{TRAIL_TOKEN}for",
+    "try_tokens": f'{TRAIL_TOKEN}"FOR"',
+}
+
+# Conditionals
+LEVEL_100: GRAMMAR_DICT = {
+    "iff": '"IFF" lpar arrangement rpar "THN" lpar press_message rpar ("ELS" lpar press_message rpar)?',
+    "press_message": f"{TRAIL_TOKEN}iff",
+    "try_tokens": f'{TRAIL_TOKEN}"IFF"',
+}
+
+# Puppets and Favors
+LEVEL_110: GRAMMAR_DICT = {
+    "xoy": '"XOY" lpar power rpar lpar power rpar',
+    "ydo": '"YDO" lpar power rpar (lpar unit rpar)+',
+    "arrangement": f"{TRAIL_TOKEN}xoy / ydo",
+    "try_tokens": f'{TRAIL_TOKEN}"XOY" / "YDO"',
+}
+
+# Forwarding Press
+LEVEL_120: GRAMMAR_DICT = {
+    "snd": '"SND" lpar power rpar lpar power (ws power)* rpar lpar message rpar',
+    "fwd": '"FWD" lpar power (ws power)* rpar lpar power rpar lpar power rpar',
+    "bcc": '"BCC" lpar power rpar lpar power (ws power)* rpar lpar power rpar',
+    "arrangement": f"{TRAIL_TOKEN}snd / fwd / bcc",
+    "try_tokens": f'{TRAIL_TOKEN}"SND" / "FWD" / "BCC"',
+}
+
+# Explanations
+LEVEL_130: GRAMMAR_DICT = {
+    "fct_thk_prp_ins": "fct / thk / prp / ins",
+    "qry_wht_prp_ins_sug": "qry / wht/ prp / ins / sug",  # added sug because it looks like it's also supported at level 130
+    "why": '"WHY" lpar fct_thk_prp_ins rpar',
+    "pob": '"POB" lpar why rpar',
+    "idk": '"IDK" lpar qry_wht_prp_ins_sug rpar',
+    "reply": f"{TRAIL_TOKEN}why / pob / idk",
+    "try_tokens": f'{TRAIL_TOKEN}"IDK" / "WHY" / "POB"',
+}
+
+LEVELS: Tuple[GRAMMAR_DICT] = (
+    LEVEL_10,
+    LEVEL_20,
+    LEVEL_30,
+    LEVEL_40,
+    LEVEL_50,
+    LEVEL_60,
+    LEVEL_70,
+    LEVEL_80,
+    LEVEL_90,
+    LEVEL_100,
+    LEVEL_110,
+    LEVEL_120,
+    LEVEL_130,
+)
+
+GRAMMAR_DICT = Dict[str, str]
+
+
+def create_daide_grammar(
+    level: DAIDE_LEVEL = 30, allow_just_arrangment: bool = False
+) -> DAIDEGrammar:
+    """Create a DAIDEGrammar object given a level of DAIDE.
+
+    The DAIDEGrammar object inherits from parsimonious.grammar.Grammar,
+    see https://github.com/erikrose/parsimonious for more information.
 
-grammar = Grammar(
+    Args:
+        level (DAIDE_LEVEL, optional): level of DIADE to create grammar for. Defaults to 30.
+        allow_just_arrangement (bool, optional): if set to True, the parser accepts strings that
+        are only arrangements, in addition to press messages. So, for example, the parser could parse
+        'PCE (GER ITA)'. Normally, this would raise a ParseError.
+
+    Returns:
+        DAIDEGrammar: Grammar object
+    """
+    grammar_str = _create_daide_grammar_str(level, allow_just_arrangment)
+    grammar = DAIDEGrammar(grammar_str)
+    return grammar
+
+
+def _create_daide_grammar_dict(level: DAIDE_LEVEL = 30) -> GRAMMAR_DICT:
+    """Combine DAIDE grammar dicts into one dict.
+
+    Args:
+        level (DAIDE_LEVEL, optional): The level of DAIDE to make grammar for. Defaults to 30.
+
+    Returns:
+        GRAMMAR_DICT: Dictionary representing all rules for passed daide level
+    """
+
+    level_idxs = list(range(int((level / 10))))
+    grammar: GRAMMAR_DICT = {}
+
+    for level_idx in level_idxs:
+        new_grammar = LEVELS[level_idx]
+        grammar = _merge_grammars(grammar, new_grammar)
+    return grammar
+
+
+def _create_daide_grammar_str(
+    level: DAIDE_LEVEL = 30, allow_just_arrangement: bool = False
+) -> str:
+    """Create string representing DAIDE grammar in PEG
+
+    Args:
+        level (DAIDE_LEVEL, optional): _description_. Defaults to 30.
+
+    Returns:
+        str: string representing DAIDE grammar in PEG
     """
-    message = press_message / reply
-    press_message = prp / ccl / fct / thk / try / ins / qry / sug / wht / how / exp / iff / frm
+    grammar_dict = _create_daide_grammar_dict(level)
+    grammar_str = _create_grammar_str_from_dict(grammar_dict, allow_just_arrangement)
+    return grammar_str
+
+
+def _create_grammar_str_from_dict(
+    grammar: GRAMMAR_DICT, allow_just_arrangement: bool = False
+) -> str:
+    grammar_str = ""
+    for item in grammar.items():
+        # message needs to be the first rule in the string, per parsimonious rules:
+        # "The first rule is taken to be the default start symbol, but you can override that."
+        # https://github.com/erikrose/parsimonious#example-usage
+        if item[0] == "message":
+            left = item[0]
+            right = item[1]
+            if allow_just_arrangement:
+                right += " / arrangement"
+            grammar_str = f"{left} = {right}\n" + grammar_str
+
+        else:
+            grammar_str += f"{item[0]} = {item[1]}\n"
+    return grammar_str
+
+
+def _merge_grammars(
+    old_grammar: GRAMMAR_DICT, new_grammar: GRAMMAR_DICT
+) -> GRAMMAR_DICT:
+    old_keys = set(old_grammar.keys())
+    new_keys = set(new_grammar.keys())
+
+    old_unique = old_keys.difference(new_keys)
+    new_unique = new_keys.difference(old_keys)
+    shared_keys = new_keys.intersection(old_keys)
+
+    merged_grammar: GRAMMAR_DICT = {}
+    for key in old_unique:
+        merged_grammar[key] = old_grammar[key]
+    for key in new_unique:
+        merged_grammar[key] = new_grammar[key]
+    for key in shared_keys:
+        merged_grammar[key] = _merge_shared_key_value(old_grammar, new_grammar, key)
+    return merged_grammar
+
+
+def _merge_shared_key_values(
+    old_grammar: GRAMMAR_DICT, new_grammar: GRAMMAR_DICT, shared_keys: Set[str]
+) -> GRAMMAR_DICT:
+    merged_grammar = {}
+    for key in shared_keys:
+        merged_grammar[key] = _merge_shared_key_value(old_grammar, new_grammar, key)
+    return merged_grammar
+
+
+def _merge_shared_key_value(
+    old_grammar: GRAMMAR_DICT, new_grammar: GRAMMAR_DICT, shared_key: str
+) -> str:
+
+    if new_grammar[shared_key][:3] == TRAIL_TOKEN:
+        new_value = old_grammar[shared_key] + " / " + new_grammar[shared_key][3:]
+    else:
+        new_value = new_grammar[shared_key]
+    return new_value
+
+
+# TODO: we need a test to confirm that this string == create_daide_grammar(130)
+FULL_DAIDE_GRAMMAR_STRING = """
+    message = press_message / reply 
+    press_message = prp / ccl / fct / thk / try / ins / qry / sug / wht / how / exp / iff / frm 
     prp = "PRP" lpar arrangement rpar 
     ins = "INS" lpar arrangement rpar
     qry = "QRY" lpar arrangement rpar
     sug = "SUG" lpar arrangement rpar
-    ccl = "CCL" lpar press_message rpar
-    fct = ("FCT" lpar qry rpar) / ("FCT" lpar "NOT" lpar qry rpar rpar) / ("FCT" lpar arrangement rpar)
+    ccl = "CCL" lpar press_message rpar 
+    fct = ("FCT" lpar qry rpar) / ("FCT" lpar "NOT" lpar qry rpar rpar) / ("FCT" lpar arrangement rpar) 
     thk = ("THK" lpar qry rpar) / ("THK" lpar "NOT" lpar qry rpar rpar) / ("THK" lpar arrangement rpar)
-    try = "TRY" lpar try_tokens (ws try_tokens)* rpar
+    try = "TRY" lpar try_tokens (ws try_tokens)* rpar 
     wht = "WHT" lpar unit rpar
-    how = "HOW" lpar province rpar
+    how = ("HOW" lpar province rpar) / ("HOW" lpar power rpar)
     exp = "EXP" lpar turn rpar lpar message rpar
     iff = "IFF" lpar arrangement rpar "THN" lpar press_message rpar ("ELS" lpar press_message rpar)?
     frm = "FRM" lpar power rpar lpar power (ws power)* rpar lpar message rpar
 
-    reply = yes / rej / bwx / huh / fct / thk / idk / sry / why / pob
-    yes = "YES" lpar press_message rpar
-    rej = "REJ" lpar press_message rpar
-    bwx = "BWX" lpar press_message rpar
+    reply = yes / rej / bwx / huh / fct / thk / idk / sry / why / pob 
+    yes = "YES" lpar press_message rpar 
+    rej = "REJ" lpar press_message rpar 
+    bwx = "BWX" lpar press_message rpar 
     huh = "HUH" lpar press_message rpar
     idk = "IDK" lpar qry_wht_prp_ins rpar
     sry = "SRY" lpar exp rpar
@@ -49,7 +372,7 @@
     fwd = "FWD" lpar power (ws power)* rpar lpar power rpar lpar power rpar
     bcc = "BCC" lpar power rpar lpar power (ws power)* rpar lpar power rpar
 
-    order = hld / mto / sup/ cvy / move_by_cvy / retreat / build
+    order = hld / mto / sup / cvy / move_by_cvy / retreat / build
     hld = lpar unit rpar "HLD"
     mto = lpar unit rpar "MTO" ws province
     sup = lpar unit rpar "SUP" lpar unit rpar ("MTO" ws province)?
@@ -79,7 +402,6 @@
     lpar = ~"\s*\(\s*"
     rpar = ~"\s*\)\s*"
     ws = ~"\s+"
-    qry_wht_prp_ins = qry / wht/ prp /ins
+    qry_wht_prp_ins_sug = qry / wht/ prp / ins / sug
     fct_thk_prp_ins = fct / thk / prp / ins
     """
-)