aboutsummaryrefslogtreecommitdiff
path: root/neb
diff options
context:
space:
mode:
Diffstat (limited to 'neb')
-rw-r--r--neb/__init__.py6
-rw-r--r--neb/exceptions.py13
-rw-r--r--neb/interpreter.py876
-rw-r--r--neb/lexer.py168
-rw-r--r--neb/parser.py67
-rw-r--r--neb/structs.py108
-rw-r--r--neb/typeclass.py34
7 files changed, 1272 insertions, 0 deletions
diff --git a/neb/__init__.py b/neb/__init__.py
new file mode 100644
index 0000000..f5afe60
--- /dev/null
+++ b/neb/__init__.py
@@ -0,0 +1,6 @@
+from .structs import *
+from .lexer import *
+from .parser import *
+from .interpreter import *
+from .exceptions import *
+from .typeclass import *
diff --git a/neb/exceptions.py b/neb/exceptions.py
new file mode 100644
index 0000000..8bbe000
--- /dev/null
+++ b/neb/exceptions.py
@@ -0,0 +1,13 @@
+class NebPanic(BaseException):
+ pass
+
+class InterpretPanic(NebPanic):
+ def __init__(self, sym, msg, arg=None):
+ big_message = f"[{sym.line}] '{sym.name}': {msg}"
+ if arg is not None:
+ big_message += f" (got {arg})"
+ super().__init__(big_message)
+
+class LexError(NebPanic):
+ def __init__(self, message, line):
+ super().__init__(f"line {line}: {message}")
diff --git a/neb/interpreter.py b/neb/interpreter.py
new file mode 100644
index 0000000..760b3a6
--- /dev/null
+++ b/neb/interpreter.py
@@ -0,0 +1,876 @@
+from .structs import *
+from .exceptions import *
+from .lexer import lex
+from .parser import parse
+from .typeclass import TypeEnum, is_subtype_of
+from pathlib import Path
+from glob import glob
+import subprocess
+import shlex
+import random
+import sys
+import math
+
+
+class Arg:
+
+ def __init__(self, name, type_, *, optional=False, lazy=False):
+ self.name = name
+ self.type_ = type_
+ self.optional = optional
+ self.lazy = lazy
+
+ def __str__(self):
+ opt = "?" if self.optional else ""
+ lazy = "~" if self.lazy else ""
+ return f"{lazy}{opt}{self.name} {self.type_}"
+
+
+class Function:
+
+ def __init__(self, name, params, body, args=None, many=None):
+ self.name = name
+ self.params = params
+ self.body = body
+ self.args = args
+ self.many = many
+ self.type_ = TypeEnum.ANY # TODO no it's not
+
+ def describe(self, name=None):
+ if name is None:
+ name = self.name
+ out = [f"({name}"]
+ if self.args is not None:
+ for arg in self.args:
+ out.append(f"{arg}")
+ if self.many is not None:
+ out.append(f"{self.many}")
+ return " ".join(out) + ")"
+
+ def arity_check(self, symbol, params):
+ min_arity = len([a for a in self.args if not a.optional])
+ max_arity = -1 if self.many is not None else len(self.args)
+
+ if len(params) < min_arity or (max_arity >= 0 and len(params) > max_arity):
+ if max_arity < 0:
+ fmt = f"{min_arity}+"
+ elif min_arity != max_arity:
+ fmt = f"{min_arity}-{max_arity}"
+ else:
+ fmt = f"{min_arity}"
+ raise InterpretPanic(symbol, f"expected [{fmt}] arguments, received {len(params)}")
+ return True
+
+ def evaluate_args(self, symbol, params, env, ns):
+ self.arity_check(symbol, params)
+ ret = []
+
+ for idx, param in enumerate(params):
+ if idx < len(self.args):
+ arg = self.args[idx]
+ else:
+ arg = self.many
+ if arg.lazy:
+ ret.append(param)
+ continue
+ ev = evaluate(param, env, ns)
+ if not is_subtype_of(ev.type_, arg.type_):
+ exp = f"{arg.type_}"
+ rec = f"{ev.type_}"
+ raise InterpretPanic(symbol, f"received {rec}, expected {exp}", ev)
+ ret.append(ev)
+ return ret
+
+ def call(self, expr, env):
+ pass
+
+class Builtin(Function):
+
+ def __init__(self, callable_, args=None, many=None):
+ super().__init__("<builtin>", None, callable_, args, many)
+
+ def call(self, expr, env, ns):
+ self.arity_check(expr.args[0], expr.args[1:])
+ evaluated_args = self.evaluate_args(expr.args[0], expr.args[1:], env, ns)
+ return self.body(expr.args[0], evaluated_args, env, ns)
+
+
+class UserFunction(Function):
+
+ def __init__(self, name, params, body):
+ newparams, args, many = self.process_params(name, params)
+ super().__init__(name, newparams, body, args, many)
+
+ def process_params(self, name, params):
+ newparams = []
+ args = []
+ many = None
+ prev_type = False
+ first = True
+ for param in params:
+ if isinstance(param, Symbol):
+ if many is not None:
+ raise NebPanic("& must be last argument")
+ if param.name == "&":
+ many = Arg(param.name, TypeEnum.ANY)
+ else:
+ newparams.append(param)
+ args.append(Arg(param.name, TypeEnum.ANY))
+ prev_type = False
+ elif isinstance(param, Type) and not prev_type and not first:
+ typ = TypeEnum.__getattr__(param.name[1:].upper())
+ if many is None:
+ args[-1].type_ = typ
+ else:
+ many.type_ = typ
+ prev_type = True
+ else:
+ raise NebPanic("invalid :func signature", param)
+ first = False
+ return newparams, args, many
+
+ def call(self, expr, env, ns):
+ self.arity_check(expr.args[0], expr.args[1:])
+ evaluated_args = self.evaluate_args(expr.args[0], expr.args[1:], env, ns)
+ this_env = Environment(env)
+ for idx, param in enumerate(self.params):
+ this_env.register(param.name, evaluated_args[idx])
+
+ # if we got "many", wrap the rest in a list
+ if self.many:
+ this_env.register(self.many.name, List(evaluated_args[len(self.params):], True))
+
+ return interpret(self.body, env=this_env, ns=ns)
+
+class Environment:
+
+ def __init__(self, parent=None):
+ self.parent = parent
+ self.environment = {}
+
+ def register(self, key, value):
+ self.environment[key] = value
+
+ def reregister(self, key, value):
+ if not self.contains(key):
+ raise NebPanic(f"undefined symbol: '{key}")
+ if key in self.environment:
+ self.register(key, value)
+ else:
+ self.parent.reregister(key, value)
+
+ def contains(self, key):
+ if key in self.environment:
+ return True
+ elif self.parent is not None:
+ return self.parent.contains(key)
+ else:
+ return False
+
+ def get(self, key):
+ if not self.contains(key):
+ raise NebPanic(f"undefined symbol: '{key}")
+ if key in self.environment:
+ return self.environment[key]
+ else:
+ return self.parent.get(key)
+
+ def __str__(self):
+ out = ""
+ for k, v in self.environment.items():
+ out += f"{k}: {v}, "
+ return out
+
+GLOBALS = Environment()
+
+def interpret(exprs, *, env=GLOBALS, ns=None):
+ ret = None
+ for expr in exprs:
+ ret = evaluate(expr, env, ns)
+ return ret
+
+def evaluate(expr, env, ns=None):
+ if isinstance(expr, Literal) or isinstance(expr, Function) or isinstance(expr, Type):
+ return expr
+ elif isinstance(expr, Symbol):
+ if env.contains(expr.name):
+ return evaluate(env.get(expr.name), env, ns)
+ elif ns is not None and env.contains(f"{ns}/{expr.name}"):
+ return evaluate(env.get(f"{ns}/{expr.name}"), env, ns)
+ else:
+ raise NebPanic(f"no such symbol: {expr}")
+
+ # if it's a literal list, return it
+ if expr.data:
+ return expr
+ # if it's an empty list, return it
+ elif len(expr.args) == 0:
+ return expr
+
+ if not isinstance(expr.args[0], Symbol):
+ raise NebPanic("can't evaluate without a symbol")
+ name = expr.args[0].name
+ if env.contains(name):
+ return env.get(name).call(expr, env, ns)
+ elif ns is not None and env.contains(f"{ns}/{name}"):
+ return env.get(f"{ns}/{name}").call(expr, env, ns)
+ else:
+ raise InterpretPanic(expr.args[0], "unable to evaluate")
+
+def interpretOr(symbol, args, env, ns):
+ # or returns true for the first expression that returns true
+ for arg in args:
+ ev = evaluate(arg, env, ns)
+ if not isinstance(ev, Bool):
+ raise InterpretPanic(symbol, "requires :bool arguments")
+ if ev.value == True:
+ return ev
+ return Bool(False)
+
+or_arg = Arg("arg", TypeEnum.BOOL, lazy=True)
+GLOBALS.register("or", Builtin(interpretOr, [or_arg, or_arg], or_arg))
+
+def interpretAnd(symbol, args, env, ns):
+ # and returns false for the first expression that returns false
+ for arg in args:
+ ev = evaluate(arg, env, ns)
+ if not isinstance(ev, Bool):
+ raise InterpretPanic(symbol, "requires :bool arguments")
+ if ev.value == False:
+ return ev
+ return Bool(True)
+
+GLOBALS.register("and", Builtin(interpretAnd, [or_arg, or_arg], or_arg))
+
+def interpretEq(symbol, args, env, ns):
+ # NOTE this currently only works for literals
+ # compare types because 0 != #false in neb
+ if type(args[0]) == type(args[1]) and args[0].value == args[1].value:
+ return Bool(True)
+ else:
+ return Bool(False)
+
+eq_arg = Arg("value", TypeEnum.LITERAL)
+GLOBALS.register("eq?", Builtin(interpretEq, [eq_arg, eq_arg]))
+
+def interpretGreaterThan(symbol, args, env, ns):
+ return Bool(args[0].value > args[1].value)
+
+compare_arg = Arg("num", TypeEnum.NUMBER)
+GLOBALS.register(">", Builtin(interpretGreaterThan, [compare_arg, compare_arg]))
+
+def interpretGreaterThanEqual(symbol, args, env, ns):
+ return Bool(args[0].value >= args[1].value)
+
+GLOBALS.register(">=", Builtin(interpretGreaterThanEqual, [compare_arg, compare_arg]))
+
+def interpretLessThan(symbol, args, env, ns):
+ return Bool(args[0].value < args[1].value)
+
+GLOBALS.register("<", Builtin(interpretLessThan, [compare_arg, compare_arg]))
+
+def interpretLessThanEqual(symbol, args, env, ns):
+ return Bool(args[0].value <= args[1].value)
+
+GLOBALS.register("<=", Builtin(interpretLessThanEqual, [compare_arg, compare_arg]))
+
+def interpretAddition(symbol, args, env, ns):
+ res = 0
+ for arg in args:
+ res += arg.value
+ if isinstance(res, float):
+ return Float(res)
+ else:
+ return Int(res)
+
+term_arg = Arg("term", TypeEnum.NUMBER)
+GLOBALS.register("+", Builtin(interpretAddition, [term_arg], term_arg))
+
+def interpretSubtraction(symbol, args, env, ns):
+ if len(args) == 1:
+ res = -args[0].value
+ else:
+ res = args[0].value
+ for arg in args[1:]:
+ res -= arg.value
+ if isinstance(res, float):
+ return Float(res)
+ else:
+ return Int(res)
+
+GLOBALS.register("-", Builtin(interpretSubtraction, [term_arg], term_arg))
+
+def interpretMultiplication(symbol, args, env, ns):
+ res = args[0].value
+ for arg in args[1:]:
+ res = res * arg.value
+ if isinstance(res, float):
+ return Float(res)
+ else:
+ return Int(res)
+
+factor_arg = Arg("factor", TypeEnum.NUMBER)
+GLOBALS.register("*", Builtin(interpretMultiplication, [factor_arg, factor_arg], factor_arg))
+
+def interpretDivision(symbol, args, env, ns):
+ ret = args[0].value / args[1].value
+ if int(ret) == ret:
+ return Int(int(ret))
+ else:
+ return Float(ret)
+
+GLOBALS.register("/", Builtin(interpretDivision, [factor_arg, factor_arg]))
+
+def interpretNot(symbol, args, env, ns):
+ return Bool(not args[0].value)
+
+not_arg = Arg("not", TypeEnum.BOOL)
+GLOBALS.register("not", Builtin(interpretNot, [not_arg]))
+
+def interpretIf(symbol, args, env, ns):
+ if args[0].value:
+ return evaluate(args[1], env, ns)
+ elif len(args) == 3:
+ return evaluate(args[2], env, ns)
+ return List([])
+
+cond = Arg("cond", TypeEnum.BOOL)
+t_branch = Arg("t-branch", TypeEnum.ANY, lazy=True)
+f_branch = Arg("f-branch", TypeEnum.ANY, optional=True, lazy=True)
+GLOBALS.register("if", Builtin(interpretIf, [cond, t_branch, f_branch]))
+
+def interpretPrint(symbol, args, env, ns):
+ print(args[0].value)
+ return List([]) # print returns nothing
+
+GLOBALS.register("print", Builtin(interpretPrint, [Arg("arg", TypeEnum.STRING)]))
+
+def interpretDef(symbol, args, env, ns):
+
+ if not isinstance(args[0], Symbol):
+ raise InterpretPanic(symbol, "requires a :string name", args[0])
+ name = args[0].name # NOTE: we are not evaluating the name!!
+ if not isinstance(name, str):
+ raise InterpretPanic(symbol, "requires a :string name")
+
+ env.register(name, args[1]) # TODO since this isn't lazily evaluated, side effects are allowed (bad!)
+
+ return List([])
+
+def_name_arg = Arg("name", TypeEnum.ANY, lazy=True)
+def_val_arg = Arg("value", TypeEnum.ANY)
+GLOBALS.register("def", Builtin(interpretDef, [def_name_arg, def_val_arg]))
+
+def interpretRedef(symbol, args, env, ns):
+ if not isinstance(args[0], Symbol):
+ raise InterpretPanic(symbol, "requires a :string name", args[0])
+ name = args[0].name # NOTE: we are not evaluating the name!!
+ if not env.contains(name):
+ raise InterpretPanic(symbol, "not previously defined", args[0])
+
+ env.reregister(name, args[1])
+ return List([])
+
+GLOBALS.register("redef", Builtin(interpretRedef, [def_name_arg, def_val_arg]))
+
+def interpretLambda(symbol, args, env, ns):
+ if len(args[0].args) != 0:
+ func = UserFunction("<lambda>", args[0].args, args[1:])
+ else:
+ func = UserFunction("<lambda>", [], args[1:])
+ return func
+
+lambda_args_arg = Arg("args", TypeEnum.ANY, lazy=True)
+lambda_body_arg = Arg("body", TypeEnum.ANY, lazy=True)
+GLOBALS.register("lambda", Builtin(interpretLambda, [lambda_args_arg, lambda_body_arg], lambda_body_arg))
+
+def interpretToString(symbol, args, env, ns):
+ item = args[0]
+ if isinstance(item, String):
+ return item
+ elif isinstance(item, Literal):
+ return String(str(item))
+ else:
+ return String(f"{item}")
+
+GLOBALS.register("->string", Builtin(interpretToString, [Arg("arg", TypeEnum.ANY)]))
+
+def interpretConcat(symbol, args, env, ns):
+ out = ""
+ for arg in args:
+ out += arg.value
+ return String(out)
+
+string_arg = Arg("arg", TypeEnum.STRING)
+GLOBALS.register("concat", Builtin(interpretConcat, [string_arg, string_arg], string_arg))
+
+def interpretForCount(symbol, args, env, ns):
+ new_env = Environment(env)
+ ret = None
+ for idx in range(0, args[0].value):
+ new_env.register("idx", Int(idx + 1))
+ for arg in args[1:]:
+ ret = evaluate(arg, new_env, ns)
+ if ret is None:
+ return List([])
+ return ret
+
+for_count_arg = Arg("count", TypeEnum.INT)
+for_body_arg = Arg("body", TypeEnum.ANY, lazy=True)
+GLOBALS.register("for-count", Builtin(interpretForCount, [for_count_arg, for_body_arg], for_body_arg))
+
+def interpretForEach(symbol, args, env, ns):
+ new_env = Environment(env)
+ ret = None
+ for item in args[0].args:
+ new_env.register("_item_", evaluate(item, env, ns))
+ for arg in args[1:]:
+ ret = evaluate(arg, new_env, ns)
+ if ret is None:
+ return List([])
+ return ret
+
+for_each_arg = Arg("list", TypeEnum.LIST)
+GLOBALS.register("for-each", Builtin(interpretForEach, [for_each_arg, for_body_arg], for_body_arg))
+
+def interpretPipe(symbol, args, env, ns):
+ new_env = Environment(env)
+ pipe = None
+ for arg in args:
+ if pipe is not None:
+ new_env.register("items", pipe)
+ pipe = evaluate(arg, new_env, ns)
+ if pipe is None:
+ return List([])
+ return pipe
+
+# TODO
+GLOBALS.register("|", Builtin(interpretPipe, 2))
+
+def interpretBranch(symbol, args, env, ns):
+ for arg in args:
+ if len(arg.args) != 2:
+ raise InterpretPanic(symbol, "each branch requires two expressions")
+ cond = evaluate(arg.args[0], env, ns) # this is the condition
+ if not isinstance(cond, Bool):
+ raise InterpretPanic(symbol, "branch condition must be :bool", cond)
+ if cond.value:
+ return evaluate(arg.args[1], env, ns)
+ return List([])
+
+GLOBALS.register("branch", Builtin(interpretBranch, [for_body_arg], for_body_arg))
+
+def interpretFunc(symbol, args, env, ns):
+ if not isinstance(args[0], Symbol):
+ raise InterpretPanic(symbol, "requires a :string name")
+ name = args[0].name # NOTE: we are not evaluating the name!!
+
+ if ns is not None:
+ name = f"{ns}/{name}"
+
+ # compose a lambda
+ func = interpretLambda(None, args[1:], env, ns)
+
+ env.register(name, func)
+ return List([])
+
+GLOBALS.register("func", Builtin(interpretFunc, [def_name_arg, lambda_args_arg, lambda_body_arg], lambda_body_arg))
+
+def interpretReadLines(symbol, args, env, ns):
+ target_file_name = args[0].value
+ target_file = Path(target_file_name).resolve()
+ if not target_file.exists():
+ raise InterpretPanic(symbol, "no such file", target_file)
+ with open(target_file, "r") as fil:
+ data = fil.readlines()
+ out = List([String(d) for d in data], True) # all lines are strings
+ return out
+
+GLOBALS.register("read-lines", Builtin(interpretReadLines, [Arg("filename", TypeEnum.STRING)]))
+
+def interpretStrip(symbol, args, env, ns):
+ return String(args[0].value.strip())
+
+GLOBALS.register("strip", Builtin(interpretStrip, [Arg("filename", TypeEnum.STRING)]))
+
+# - string->int and string->float
+def interpretStringToInt(symbol, args, env, ns):
+ try:
+ val = int(args[0].value)
+ return Int(val)
+ except:
+ raise InterpretPanic(symbol, "can't convert to an :int", args[0])
+
+GLOBALS.register("string->int", Builtin(interpretStringToInt, [Arg("arg", TypeEnum.STRING)]))
+
+def interpretSplit(symbol, args, env, ns):
+ target = args[0]
+ if len(args) == 1:
+ return List([String(char) for char in target.value], True)
+ splitter = args[1]
+ ret = target.value.split(splitter.value)
+ return List([String(r) for r in ret], True)
+
+GLOBALS.register("split", Builtin(interpretSplit, [Arg("target", TypeEnum.STRING)], Arg("splitter", TypeEnum.STRING, optional=True)))
+
+def interpretListLength(symbol, args, env, ns):
+ return Int(len(args[0].args))
+
+GLOBALS.register("list-length", Builtin(interpretListLength, [Arg("arg", TypeEnum.LIST)]))
+
+def interpretFirst(symbol, args, env, ns):
+ if len(args[0].args) == 0:
+ raise InterpretPanic(symbol, "list is empty")
+ return evaluate(args[0].args[0], env, ns)
+
+GLOBALS.register("first", Builtin(interpretFirst, [Arg("arg", TypeEnum.LIST, )]))
+
+def interpretRest(symbol, args, env, ns):
+ # TODO do we know it's not evaluated?
+ return List(args[0].args[1:], True) # we don't evaluate the remainder of the list
+
+GLOBALS.register("rest", Builtin(interpretRest, [Arg("arg", TypeEnum.LIST)]))
+
+def interpretMap(symbol, args, env, ns):
+ func = args[0]
+ if not isinstance(func, Function):
+ raise InterpretPanic(symbol, "requires a :func as its first argument", func)
+ lst = evaluate(args[1], env, ns)
+ if not isinstance(lst, List):
+ raise InterpretPanic(symbol, "requires a :list as its second argument", lst)
+ out = []
+ for arg in lst.args:
+ ev = func.call(List([func, arg]), env, ns)
+ out.append(ev)
+ return List(out, True)
+
+GLOBALS.register("map", Builtin(interpretMap, [Arg("func", TypeEnum.ANY), Arg("list", TypeEnum.LIST)]))
+
+def interpretZip(symbol, args, env, ns):
+ z1 = args[0]
+ z2 = args[1]
+ if len(z1.args) != len(z2.args):
+ raise InterpretPanic(symbol, "requires two :lists of the same size")
+ out = []
+ for idx in range(len(z1.args)):
+ f = z1.args[idx]
+ s = z2.args[idx]
+ out.append(List([f, s], True))
+ return List(out, True)
+
+zip_arg = Arg("list", TypeEnum.LIST)
+GLOBALS.register("zip", Builtin(interpretZip, [zip_arg, zip_arg]))
+
+def interpretList(symbol, args, env, ns):
+ return List(args, True)
+
+GLOBALS.register("list", Builtin(interpretList, [], Arg("item", TypeEnum.ANY)))
+
+def interpretListReverse(symbol, args, env, ns):
+ new_args = args[0].args[:] # make a copy of the args
+ new_args.reverse()
+ return List(new_args, True)
+
+GLOBALS.register("list-reverse", Builtin(interpretListReverse, [Arg("list", TypeEnum.LIST)]))
+
+def interpretApply(symbol, args, env, ns):
+ # TODO: to support lambdas, we can't assume the func is defined
+ func = args[0]
+ if not isinstance(func, Symbol):
+ raise InterpretPanic(symbol, "requires a symbol as its first argument", func)
+ new_lst = List([func] + args[1].args)
+ return evaluate(new_lst, env, ns)
+
+GLOBALS.register("apply", Builtin(interpretApply, [Arg("func", TypeEnum.ANY, lazy=True), Arg("list", TypeEnum.LIST)]))
+
+def interpretGlob(symbol, args, env, ns):
+ items = glob(args[0].value)
+ return List([String(item) for item in items], True)
+
+GLOBALS.register("glob", Builtin(interpretGlob, [Arg("regex", TypeEnum.STRING)]))
+
+def interpretShell(symbol, args, env, ns):
+ ret = subprocess.run(shlex.split(args[0].value), capture_output=True)
+ return List([String(r) for r in ret.stdout.decode("utf-8").split("\n")], True)
+
+GLOBALS.register("$", Builtin(interpretShell, [Arg("command", TypeEnum.STRING)]))
+
+def interpretEmpty(symbol, args, env, ns):
+ return Bool(len(args[0].args) == 0)
+
+GLOBALS.register("empty?", Builtin(interpretEmpty, [Arg("list", TypeEnum.LIST)]))
+
+def interpretShuf(symbol, args, env, ns):
+ items = args[0].args[:]
+ random.shuffle(items)
+ return List(items, True)
+
+GLOBALS.register("shuf", Builtin(interpretShuf, [Arg("list", TypeEnum.LIST)]))
+
+def interpretIsList(symbol, args, env, ns):
+ return Bool(isinstance(args[0], List))
+
+GLOBALS.register("list?", Builtin(interpretIsList, [Arg("arg", TypeEnum.ANY)]))
+
+def interpretBlock(symbol, args, env, ns):
+ ret = List([])
+ for arg in args:
+ ret = evaluate(arg, env, ns)
+ return ret
+
+block_arg = Arg("expr", TypeEnum.ANY, lazy=True)
+GLOBALS.register("block", Builtin(interpretBlock, [block_arg], block_arg))
+
+def interpretExit(symbol, args, env, ns):
+ status = 0 if len(args) == 0 else args[0].value
+ sys.exit(status)
+ return List([])
+
+exit_arg = Arg("status", TypeEnum.INT, optional=True)
+GLOBALS.register("exit", Builtin(interpretExit, [exit_arg]))
+
+def interpretUnlink(symbol, args, env, ns):
+ target_path = Path(args[0].value).resolve()
+ if not target_path.exists():
+ raise InterpretPanic(symbol, "target file does not exist", target_path)
+ target_path.unlink()
+ return List([])
+
+GLOBALS.register("unlink", Builtin(interpretUnlink, [Arg("filename", TypeEnum.STRING)]))
+
+def interpretArgv(symbol, args, env, ns):
+ out = []
+ for arg in sys.argv[1:]:
+ out.append(String(arg))
+ return List(out, True)
+
+GLOBALS.register("argv", Builtin(interpretArgv, []))
+
+def interpretIn(symbol, args, env, ns):
+ target = args[0]
+ lst = args[1]
+ for arg in lst.args:
+ if type(arg) == type(target) and arg.value == target.value:
+ return Bool(True)
+ return Bool(False)
+
+in_target_arg = Arg("target", TypeEnum.LITERAL)
+in_list_arg = Arg("list", TypeEnum.LIST)
+GLOBALS.register("in?", Builtin(interpretIn, [in_target_arg, in_list_arg]))
+
+def interpretLast(symbol, args, env, ns):
+ if len(args[0].args) == 0:
+ raise InterpretPanic("List is empty")
+ return evaluate(args[0].args[-1], env, ns)
+
+GLOBALS.register("last", Builtin(interpretLast, [Arg("list", TypeEnum.LIST)]))
+
+def interpretJoin(symbol, args, env, ns):
+ lst = args[0]
+ target = args[1]
+ return String(target.value.join([a.value for a in lst.args]))
+
+join_list_arg = Arg("list", TypeEnum.LIST)
+join_string_arg = Arg("joiner", TypeEnum.STRING)
+GLOBALS.register("join", Builtin(interpretJoin, [join_list_arg, join_string_arg]))
+
+def interpretWithWrite(symbol, args, env, ns):
+ target_file = args[0]
+ new_env = Environment(env)
+ target_path = Path(target_file.value).resolve()
+ ret = Literal([])
+ with open(str(target_path), "w") as fil:
+ new_env.register("_file_", List([fil], True)) # TODO wrong!
+ for arg in args[1:]:
+ ret = evaluate(arg, new_env, ns)
+ return ret
+
+GLOBALS.register("with-write", Builtin(interpretWithWrite, [Arg("filename", TypeEnum.STRING)], Arg("exprs", TypeEnum.ANY, lazy=True)))
+
+def interpretWrite(symbol, args, env, ns):
+ # write :string :filehandle
+ line = args[0]
+ handle = args[1]
+ handle.args[0].write(line.value) # TODO wrong! how do we evaluate a handle?
+ return Literal([])
+
+GLOBALS.register("write", Builtin(interpretWrite, [Arg("string", TypeEnum.STRING), Arg("filename", TypeEnum.LIST)]))
+
+def interpretNewline(symbol, args, env, ns):
+ return String("\n")
+
+GLOBALS.register("newline", Builtin(interpretNewline, []))
+
+def interpretExists(symbol, args, env, ns):
+ return Bool(Path(args[0].value).resolve().exists())
+
+GLOBALS.register("exists?", Builtin(interpretExists, [Arg("filename", TypeEnum.STRING)]))
+
+def interpretFirstChar(symbol, args, env, ns):
+ if len(args[0].value) == 0:
+ raise InterpretPanic(symbol, ":string is empty", ev)
+ return String(args[0].value[0])
+
+GLOBALS.register("first-char", Builtin(interpretFirstChar, [Arg("string", TypeEnum.STRING)]))
+
+def interpretRestChar(symbol, args, env, ns):
+ return String(args[0].value[1:])
+
+GLOBALS.register("rest-char", Builtin(interpretRestChar, [Arg("string", TypeEnum.STRING)]))
+
+def interpretSlice(symbol, args, env, ns):
+ lst = args[0]
+ idx = args[1]
+ if len(args) == 2:
+ return List(lst.args[idx.value - 1:])
+ length = args[2]
+ diff = idx.value - 1 + length.value
+ return List(lst.args[idx.value - 1:diff])
+
+slice_list_arg = Arg("list", TypeEnum.LIST)
+slice_idx_arg = Arg("idx", TypeEnum.INT)
+slice_length_arg = Arg("length", TypeEnum.INT, optional=True)
+GLOBALS.register("slice", Builtin(interpretSlice, [slice_list_arg, slice_idx_arg, slice_length_arg]))
+
+def interpretClear(symbol, args, env, ns):
+ subprocess.run(["clear"])
+ return List([])
+
+GLOBALS.register("clear", Builtin(interpretClear, []))
+
+def interpretReadLine(symbol, args, env, ns):
+ ret = input(args[0].value)
+ return String(ret)
+
+GLOBALS.register("read-line", Builtin(interpretReadLine, [Arg("prompt", TypeEnum.STRING)]))
+
+def interpretReadChar(symbol, args, env, ns):
+ import termios, tty
+ fd = sys.stdin.fileno()
+ old = termios.tcgetattr(fd)
+ try:
+ tty.setraw(fd)
+ ch = sys.stdin.buffer.read1(4) # some keys are >1 bytes
+ except Exception:
+ raise
+ finally:
+ termios.tcsetattr(fd, termios.TCSADRAIN, old)
+ return String(ch.decode("utf-8"))
+
+GLOBALS.register("read-char", Builtin(interpretReadChar, []))
+
+def interpretAppend(symbol, args, env, ns):
+ lst = args[0]
+ val = args[1]
+ items = lst.args[:]
+ return List(items + [val], True)
+
+GLOBALS.register("append", Builtin(interpretAppend, [Arg("list", TypeEnum.LIST), Arg("item", TypeEnum.ANY)]))
+
+# TODO: this is actually for records/structs/whatever they're called
+def interpretRemove(symbol, args, env, ns):
+ lst = args[0]
+ key = args[1]
+ out = []
+ for arg in lst.args:
+ if arg.args[0].value != key.value:
+ out.append(arg)
+ return List(out, True)
+
+GLOBALS.register("remove", Builtin(interpretRemove, [Arg("list", TypeEnum.LIST), Arg("key", TypeEnum.ANY)]))
+
+def interpretWhile(symbol, args, env, ns):
+ cond = args[0]
+ ret = List([])
+ while True:
+ ev = evaluate(cond, env, ns)
+ if not isinstance(ev, Bool):
+ raise InterpretPanic(symbol, "expects a :bool condition", ev)
+ if not ev.value:
+ break
+ for arg in args[1:]:
+ ret = evaluate(arg, env, ns)
+ return ret
+
+GLOBALS.register("while", Builtin(interpretWhile, [Arg("cond", TypeEnum.BOOL, lazy=True)], Arg("expr", TypeEnum.ANY, lazy=True)))
+
+def interpretUse(symbol, args, env, ns):
+ target_file_name = args[0].value
+ target_file = Path(target_file_name).resolve()
+ if not target_file.exists():
+ raise InterpretPanic(symbol, "no such file", target_file)
+ with open(target_file, "r") as fil:
+ data = fil.read()
+ interpret(parse(lex(data)))
+ return List([])
+
+GLOBALS.register("use", Builtin(interpretUse, [Arg("filename", TypeEnum.STRING)]))
+
+def interpretAssert(symbol, args, env, ns):
+ if args[0].value != True:
+ raise InterpretPanic(symbol, "assertion failed")
+ return List([])
+
+GLOBALS.register("assert", Builtin(interpretAssert, [Arg("cond", TypeEnum.BOOL)]))
+
+def interpretHowTo(symbol, args, env, ns):
+ if not isinstance(args[0], Symbol):
+ raise InterpretPanic(symbol, "expects a symbol", args[0])
+ sym = env.get(args[0].name)
+ print(sym.describe(args[0].name))
+ return List([])
+
+GLOBALS.register("howto", Builtin(interpretHowTo, [Arg("symbol", TypeEnum.ANY, lazy=True)]))
+
+def interpretSymbols(symbol, args, env, ns):
+ keys = list(env.environment.keys())
+ keys.sort()
+ out = ""
+ for idx, key in enumerate(keys):
+ if idx % 6 == 0:
+ print(out)
+ out = f"{key}"
+ else:
+ out = f"{out} {key}"
+ print(out)
+ return List([])
+
+GLOBALS.register("symbols", Builtin(interpretSymbols, []))
+
+def interpretUseAs(symbol, args, env, ns):
+ target_file_name = args[0].value
+ target_file = Path(target_file_name).resolve()
+ if not target_file.exists():
+ raise InterpretPanic(symbol, "no such file", target_file)
+ with open(target_file, "r") as fil:
+ data = fil.read()
+ interpret(parse(lex(data)), ns=args[1].name)
+ return List([])
+
+GLOBALS.register("use-as", Builtin(interpretUseAs, [Arg("filename", TypeEnum.STRING), Arg("namespace", TypeEnum.ANY, lazy=True)]))
+
+def interpretFloor(symbol, args, env, ns):
+ return Int(math.floor(args[0].value))
+
+GLOBALS.register("floor", Builtin(interpretFloor, [Arg("floor", TypeEnum.NUMBER)]))
+
+def interpretFilter(symbol, args, env, ns):
+ func = args[0]
+ if not isinstance(func, Function):
+ raise InterpretPanic(symbol, "requires a :func as its first argument", func)
+ lst = args[1]
+ out = []
+ for arg in lst.args:
+ ev = func.call(List([func, arg]), env, ns)
+ if not isinstance(ev, Bool):
+ raise InterpretPanic(symbol, "function must return :bool", ev)
+ if ev.value:
+ out.append(arg)
+ return List(out, True)
+
+GLOBALS.register("filter", Builtin(interpretFilter, [Arg("func", TypeEnum.ANY), Arg("list", TypeEnum.LIST)]))
+
+def interpretTypeOf(symbol, args, env, ns):
+ return Type(f"{args[0].type_}")
+
+GLOBALS.register("typeof", Builtin(interpretTypeOf, [Arg("candidate", TypeEnum.ANY)]))
diff --git a/neb/lexer.py b/neb/lexer.py
new file mode 100644
index 0000000..b522460
--- /dev/null
+++ b/neb/lexer.py
@@ -0,0 +1,168 @@
+from .structs import TokenType, Token
+from .exceptions import LexError
+import sys
+
+
+types = {
+ ":int": TokenType.INT_TYPE,
+ ":float": TokenType.FLOAT_TYPE,
+ ":number": TokenType.NUMBER_TYPE,
+ ":string": TokenType.STRING_TYPE,
+ ":list": TokenType.LIST_TYPE,
+ ":any": TokenType.ANY_TYPE,
+ ":literal": TokenType.LITERAL_TYPE,
+ ":bool": TokenType.BOOL_TYPE }
+
+keywords = {
+ "if": TokenType.IF,
+ "for-count": TokenType.FOR_COUNT,
+ "def": TokenType.DEF,
+ "lambda": TokenType.LAMBDA,
+ "&": TokenType.MANY,
+ "func": TokenType.FUNC }
+
+
+WHITESPACE = [" ", "\n", "\t"]
+SEPARATORS = WHITESPACE + [")"]
+DIGITS = list("0123456789")
+
+def lex(data):
+ start = 0
+ current = 0
+ line = 1
+ end = len(data)
+
+ tokens = []
+ while current < end:
+ char = data[current]
+ if char == ";":
+ while char != "\n" and current < end:
+ current += 1
+ char = data[current]
+ continue
+ if char == "\n":
+ line += 1
+ if char in WHITESPACE:
+ current += 1
+ continue
+ elif char == "(":
+ tokens.append(Token(TokenType.OPEN_PAREN, "(", None, line))
+ elif char == ")":
+ tokens.append(Token(TokenType.CLOSE_PAREN, ")", None, line))
+ # numbers
+ elif char in DIGITS or char == ".":
+ tok, length = get_number(data[current:], line)
+ tokens.append(tok)
+ current += length
+ # strings
+ elif char == '"':
+ tok, length, offset = get_string(data[current+1:], line)
+ tokens.append(tok)
+ current += length
+ line += offset
+ # bools
+ elif char == "#":
+ tok, length = get_bool(data[current+1:], line)
+ tokens.append(tok)
+ current += length
+ #types
+ elif char == ":":
+ tok, length = get_type(data[current:], line) # include :
+ tokens.append(tok)
+ current += length
+ # symbols
+ else:
+ tok, length = get_symbol(data[current:], line)
+ if tok.text in keywords:
+ tok.type_ = keywords[tok.text]
+ tokens.append(tok)
+ current += length
+
+ current += 1
+ tokens.append(Token(TokenType.EOF, "", None, line))
+ return tokens
+
+def get_number(data, line):
+ counter = 0
+ value = ""
+ is_float = False
+ char = data[counter]
+ while char not in SEPARATORS:
+ if char in DIGITS:
+ value += char
+ elif char == ".":
+ if is_float:
+ raise LexError("too many '.' in number", line)
+ is_float = True
+ value += char
+ else:
+ raise Exception(f"invalid number: {value}")
+ counter += 1
+ if counter >= len(data):
+ break
+ char = data[counter]
+ if is_float:
+ return Token(TokenType.FLOAT, value, float(value), line), counter - 1
+ else:
+ return Token(TokenType.INT, value, int(value), line), counter - 1
+
+
+def get_string(data, line):
+ offset = 0
+ counter = 0
+ string = ""
+ while data[counter] != '"':
+ if data[counter] == "\n":
+ offset += 1
+
+ # look ahead to see if it's a double quote
+ if data[counter] == "\\" and \
+ len(data) > counter and \
+ data[counter+1] == '"':
+ string += '"'
+ counter += 1
+ else:
+ string += data[counter]
+ counter += 1
+ if counter >= len(data):
+ raise Exception("couldn't parse string")
+ string = string.encode().decode("unicode_escape")
+ return Token(TokenType.STRING, str(string), str(string), line), counter + 1, offset
+
+def get_bool(data, line):
+ counter = 0
+ value = ""
+ while data[counter] not in SEPARATORS:
+ value += data[counter]
+ counter += 1
+ if counter >= len(data):
+ break
+ if value == "true":
+ return Token(TokenType.TRUE, "#true", True, line), 4
+ elif value == "false":
+ return Token(TokenType.FALSE, "#false", False, line), 5
+ else:
+ raise LexError("couldn't parse boolean", line)
+
+def get_symbol(data, line):
+ counter = 0
+ value = ""
+ while data[counter] not in SEPARATORS:
+ value += data[counter]
+ counter += 1
+ if counter >= len(data):
+ break
+ return Token(TokenType.SYMBOL, value, None, line), counter - 1
+
+def get_type(data, line):
+ counter = 0
+ value = ""
+ while data[counter] not in SEPARATORS:
+ value += data[counter]
+ counter += 1
+ if counter >= len(data):
+ break
+ if value not in types:
+ raise LexError(f"unrecognized type {value}", line)
+ return Token(types[value], value, None, line), counter - 1
+
diff --git a/neb/parser.py b/neb/parser.py
new file mode 100644
index 0000000..ea875fb
--- /dev/null
+++ b/neb/parser.py
@@ -0,0 +1,67 @@
+from .structs import *
+
+def parseExpression(token, prev, tokens):
+ idx = 0
+ args = []
+ prev = token
+ while tokens[idx].type_ != TokenType.CLOSE_PAREN:
+ token = tokens[idx]
+ inc = 1
+ if token.type_ == TokenType.OPEN_PAREN:
+ expr, inc = parseExpression(token, prev, tokens[idx+1:])
+ args.append(expr)
+ elif token.type_ in (TokenType.STRING, TokenType.TRUE, TokenType.FALSE, TokenType.INT, TokenType.FLOAT):
+ expr, inc = parseLiteral(token, prev, tokens[idx+1:])
+ args.append(expr)
+ elif token.type_ in (TokenType.INT_TYPE, TokenType.FLOAT_TYPE, TokenType.STRING_TYPE, TokenType.ANY_TYPE, TokenType.LIST_TYPE, TokenType.NUMBER_TYPE, TokenType.BOOL_TYPE, TokenType.LITERAL_TYPE):
+ expr, inc = parseType(token, prev, tokens[idx+1:])
+ args.append(expr)
+ else:
+ expr, inc = parseSymbol(token, prev, tokens[idx+1:])
+ args.append(expr)
+ idx += inc
+ prev = token
+
+ return List(args), idx + 2 # parens
+
+def parseSymbol(token, prev, tokens):
+ return Symbol(token.text, token.line), 1
+
+def parseLiteral(token, prev, tokens):
+ if token.type_ == TokenType.STRING:
+ return String(token.value), 1
+ elif token.type_ == TokenType.INT:
+ return Int(token.value), 1
+ elif token.type_ == TokenType.FLOAT:
+ return Float(token.value), 1
+ elif token.type_ in (TokenType.TRUE, TokenType.FALSE):
+ return Bool(token.value), 1
+ else:
+ return Literal(token.value), 1
+
+def parseType(token, prev, tokens):
+ return Type(token.text), 1
+
+def parse(tokens):
+ idx = 0
+ prev = None
+ exprs = []
+ while tokens[idx].type_ != TokenType.EOF:
+ token = tokens[idx]
+ counter = 1
+ if token.type_ == TokenType.OPEN_PAREN:
+ expr, counter = parseExpression(token, prev, tokens[idx+1:])
+ exprs.append(expr)
+ elif token.type_ in (TokenType.FALSE, TokenType.TRUE, TokenType.STRING, TokenType.INT, TokenType.FLOAT):
+ lit, counter = parseLiteral(token, prev, tokens[idx+1:])
+ exprs.append(lit)
+ elif token.type_ in (TokenType.INT_TYPE, TokenType.FLOAT_TYPE, TokenType.STRING_TYPE, TokenType.ANY_TYPE, TokenType.LIST_TYPE, TokenType.NUMBER_TYPE, TokenType.BOOL_TYPE, TokenType.LITERAL_TYPE):
+ typ, counter = parseType(token, prev, tokens[idx+1:])
+ exprs.append(typ)
+ else:
+ sym, counter = parseSymbol(token, prev, tokens[idx+1:])
+ exprs.append(sym)
+
+ idx += counter
+ prev = token
+ return exprs
diff --git a/neb/structs.py b/neb/structs.py
new file mode 100644
index 0000000..09807db
--- /dev/null
+++ b/neb/structs.py
@@ -0,0 +1,108 @@
+from dataclasses import dataclass
+from enum import Enum, auto
+from typing import Any
+from .typeclass import TypeEnum
+
+# tokens and types
+# NOTE: this can probably be simplified
+class TokenType(Enum):
+
+ OPEN_PAREN = auto()
+ CLOSE_PAREN = auto()
+
+ EOF = auto()
+
+ # literals
+ INT = auto()
+ FLOAT = auto()
+ STRING = auto()
+ TRUE = auto()
+ FALSE = auto()
+
+ # keywords
+ IF = auto()
+ FOR_COUNT = auto()
+ DEF = auto()
+ LAMBDA = auto()
+ FUNC = auto()
+
+ # symbols
+ SYMBOL = auto()
+
+ # types
+ INT_TYPE = auto()
+ FLOAT_TYPE = auto()
+ NUMBER_TYPE = auto()
+ STRING_TYPE = auto()
+ ANY_TYPE = auto()
+ LIST_TYPE = auto()
+ LITERAL_TYPE = auto()
+ BOOL_TYPE = auto()
+
+ MANY = auto()
+
+@dataclass
+class Token:
+ type_: TokenType
+ text: str
+ value: Any
+ line: int
+
+ def __str__(self):
+ return f"{self.type_.name} {self.text} {self.line}"
+
+class Literal:
+ def __init__(self, value, type_=None):
+ self.value = value
+ if type_ is None:
+ self.type_ = TypeEnum.ANY
+ else:
+ self.type_ = type_
+ def __str__(self):
+ return f"{self.value}:literal"
+
+class Int(Literal):
+ def __init__(self, value):
+ super().__init__(value, TypeEnum.INT)
+ def __str__(self):
+ return f"{self.value}"
+
+class Float(Literal):
+ def __init__(self, value):
+ super().__init__(value, TypeEnum.FLOAT)
+ def __str__(self):
+ return f"{self.value}"
+
+class Bool(Literal):
+ def __init__(self, value):
+ super().__init__(value, TypeEnum.BOOL)
+ def __str__(self):
+ return f"#{str(self.value).lower()}"
+
+class String(Literal):
+ def __init__(self, value):
+ super().__init__(value, TypeEnum.STRING)
+ def __str__(self):
+ return f'"{repr(self.value)[1:-1]}"'
+
+class Type:
+ def __init__(self, name):
+ self.name = name
+ def __str__(self):
+ return self.name
+
+class Symbol:
+ def __init__(self, name, line):
+ self.name = name
+ self.line = line
+ def __str__(self):
+ return f"'{self.name}"
+
+class List:
+ def __init__(self, args, data=False):
+ self.args = args
+ self.data = data
+ self.type_ = TypeEnum.LIST
+ def __str__(self):
+ return "(" + " ".join(f"{arg}" for arg in self.args) + ")"
+
diff --git a/neb/typeclass.py b/neb/typeclass.py
new file mode 100644
index 0000000..eae412c
--- /dev/null
+++ b/neb/typeclass.py
@@ -0,0 +1,34 @@
+from enum import Enum, auto
+
+class TypeEnum(Enum):
+ ANY = auto()
+ STRING = auto()
+ INT = auto()
+ FLOAT = auto()
+ NUMBER = auto()
+ LIST = auto()
+ LITERAL = auto()
+ BOOL = auto()
+
+ def __str__(self):
+ return f":{self.name.lower()}"
+
+HIERARCHY = { TypeEnum.ANY: None,
+ TypeEnum.LITERAL: TypeEnum.ANY,
+ TypeEnum.LIST: TypeEnum.ANY,
+ TypeEnum.STRING: TypeEnum.LITERAL,
+ TypeEnum.BOOL: TypeEnum.LITERAL,
+ TypeEnum.NUMBER: TypeEnum.LITERAL,
+ TypeEnum.INT: TypeEnum.NUMBER,
+ TypeEnum.FLOAT: TypeEnum.NUMBER }
+
+def is_subtype_of(candidate, expected):
+ if candidate == expected:
+ return True
+ parent = HIERARCHY[candidate]
+ while parent is not None:
+ if parent == expected:
+ return True
+ parent = HIERARCHY[parent]
+ return False
+
generated by cgit v1.2.3 (git 2.39.1) at 2025-10-30 08:09:44 +0000