import json import re import sys tokenpatterns = [ (r’#[^\n]‘, None), (r’\s+‘, None),…
- User
import json import re import sys
token_patterns = [ (r’#[^\n]‘, None), (r’\s+‘, None), (r’[‘, ‘LBRACKET’), (r’]‘, ‘RBRACKET’), (r’{‘, ‘LBRACE’), (r’}‘, ‘RBRACE’), (r’%{‘, ‘LMAPBRACE’), (r’=>‘, ‘ARROW’), (r’,‘, ‘COMMA’), (r’(true|false)‘, ‘BOOLEAN’), (r’(0|[1-9][0-9_])‘, ‘INTEGER’), (r’:[A-Za-z_]\w*‘, ‘ATOM’), (r’[A-Za-z_]\w*‘, ‘KEY’), ]
regex_parts = [f’(?P<{name}>{pattern})’ if name else f’(?:{pattern})’ for pattern, name in token_patterns] token_regex = ‘|’.join(regex_parts) token_re = re.compile(token_regex)
class TokenizerException(Exception): pass
class ParserException(Exception): pass
def tokenize(text): pos = 0 while pos < len(text): match = token_re.match(text, pos) if not match: raise TokenizerException(f’Illegal character {text[pos]!r} at index {pos}‘) pos = match.end() if match.lastgroup: yield match.lastgroup, match.group(match.lastgroup)
class Node: def init(self, kind, value): self.kind = kind self.value = value
def to_json(self): if self.kind == “map”: return {“%k”: self.kind, “%v”: {k.to_json()[“%v”]: v.to_json() for k, v in self.value}} elif self.kind == “list”: return {“%k”: self.kind, “%v”: [item.to_json() for item in self.value]} else: return {“%k”: self.kind, “%v”: self.value}
class Parser: def init(self, tokens): self.tokens = iter(tokens) self.current_token = None self.next_token()
def next_token(self): try: self.current_token = next(self.tokens) except StopIteration: self.current_token = None
def parse(self): result = self.parse_sentence() if self.current_token is not None: raise ParserException(‘Unexpected token at the end’) return result
def parse_sentence(self): nodes = [] while self.current_token is not None: node = self.parse_data_literal() nodes.append(node) if self.current_token and self.current_token[0] in {‘COMMA’, ‘COLON’, ‘ARROW’}: self.next_token() return nodes
def parse_data_literal(self): if self.current_token[0] == ‘LBRACKET’: return self.parse_list() elif self.current_token[0] == ‘LBRACE’: return self.parse_tuple() elif self.current_token[0] == ‘LMAPBRACE’: return self.parse_map() elif self.current_token[0] == ‘INTEGER’: value = int(self.current_token[1]) self.next_token() return Node(‘int’, value) elif self.current_token[0] == ‘ATOM’: value = self.current_token[1] self.next_token() return Node(‘atom’, value) elif self.current_token[0] == ‘BOOLEAN’: value = self.current_token[1] == ‘true’ self.next_token() return Node(‘bool’, value) elif self.current_token[0] == ‘KEY’: value = self.current_token[1] self.next_token() if self.current_token and self.current_token[0] == ‘COLON’: self.next_token() return Node(‘key’, value) return Node(‘string’, value) else: raise ParserException(f’Unexpected token {self.current_token[1]}’)
def parse_list(self): self.next_token() items = [] while self.current_token and self.current_token[0] != ‘RBRACKET’: item = self.parse_data_literal() items.append(item) if self.current_token and self.current_token[0] == ‘COMMA’: self.next_token() if not self.current_token or self.current_token[0] != ‘RBRACKET’: raise ParserException(‘List not properly terminated with ]’) self.next_token() return Node(‘list’, items)
def parse_tuple(self): items = [] self.next_token() while self.current_token is not None and self.current_token[0] != ‘RBRACE’: items.append(self.parse_data_literal()) if self.current_token and self.current_token[0] == ‘COMMA’: self.next_token() if self.current_token is None or self.current_token[0] != ‘RBRACE’: raise ParserException(‘Tuple not properly terminated with }’) self.next_token() return Node(‘tuple’, [item.to_json() for item in items])
def parse_map(self): self.next_token() map_items = {} while self.current_token and self.current_token[0] != ‘RBRACE’:
if self.current_token[0] not in {‘ATOM’, ‘KEY’}: raise ParserException(f’Expected map key, found {self.current_token}‘)
key = self.current_token[1] if self.current_token[0] == ‘ATOM’ else ‘:{}’.format(self.current_token[1]) self.next_token()
if self.current_token and self.current_token[0] not in {‘ARROW’, ‘COLON’}: raise ParserException(f’Expected “=>” or “:”, found {self.current_token}’) self.next_token()
value = self.parse_data_literal() map_items[key] = value if self.current_token and self.current_token[0] == ‘COMMA’: self.next_token() if not self.current_token or self.current_token[0] != ‘RBRACE’: raise ParserException(‘Map not properly terminated with }’) self.next_token() return Node(‘map’, map_items)
def main(): try: text = sys.stdin.read()
tokens = list(tokenize(text))
if not tokens: print(json.dumps([])) else: parser = Parser(tokens) result = parser.parse()
json_output = json.dumps([node.to_json() for node in result], separators=(‘,’, ‘:’)) print(json_output) except TokenizerException as e: sys.stderr.write(str(e) + ‘\n’) except ParserException as e: sys.stderr.write(str(e) + ‘\n’)
if name == ‘main’: main() The program above is elixir-data.py program in the prj1-sol directory. I will specifically need to submit a zip-archive which unpacks into a prj1-sol directory minimally containing a file elixir-data.ebnf plus two shell scripts make.sh and run.sh: Running make.sh from any directory should build any artifacts needed to run program within the prj1-sol directory. Running run.sh from any directory should read and parse a sentence of the above language from standard input and output on staingle line containing the JSON representation of the parse. let’s write the make.sh script, which in this case may be straightforward since the core of your program is a Python script (which does not require compilation). However, the make.sh script should ensure that the project is structured correctly. i want the make.sh file to be like #!/bin/sh
#sets dir to directory containing this script dir=
dirname $0#use $dir to access programs in this directory #so that this script can be run from any directory.
echo "no build done"
and the run.sh file to be like #!/bin/sh
#sets dir to directory containing this script scriptpath=
realpath $0dir=dirname $scriptpath#use dir/ as prefix to run any programs in this dir #so that this script can be run from any directory. python3 dir/arith.py I have written my program in vscode so write the programs and commands accordingly
- Assistant
Certainly! Below are the
make.shandrun.shscripts as you have requested. Themake.shscript does not do much since Python is an interpreted language and doesn't need to be compiled, but it's implemented here to satisfy the project structure requirements.Here's what you'll want inside your
make.shscript:#!/bin/sh # Sets dir to the directory containing this script dir=$(dirname "$0") # Use $dir to access programs in this directory # so that this script can be run from any directory. echo "no build done"And for the
run.shscript, there's a slight mistake in your description—the Python script name you provided isarith.py, but based on the context, it should beelixir-data.py. I'll assume it should beelixir-data.pyand write therun.shscript accordingly:#!/bin/sh # Sets dir to the directory containing this script scriptpath=$(realpath "$0") dir=$(dirname "$scriptpath") # Use $dir/ as a prefix to run any programs in this dir # so that this script can be run from any directory. python3 "$dir/elixir-data.py"In order to use these scripts, do the following:
- Save each script in the respective
make.shandrun.shfiles in theprj1-soldirectory. - Make sure the scripts are executable. You can do this by running
chmod +x make.sh run.shfrom the terminal in theprj1-soldirectory. - Ensure that
elixir-data.pyis also in the sameprj1-soldirectory and it's executable (chmod +x elixir-data.py).
After that, executing
./make.shwon't do much but echo "no build done".When you execute
./run.shand provide input through standard input (stdin),elixir-data.pywill be invoked, process the input, and output the JSON representation of the parsed data on the command line, as your Python script dictates.Just remember to change
arith.pytoelixir-data.pyin therun.shscript if that's the correct name of the script you need to run. If you use Visual Studio Code, you can just create these.shfiles in your project folderprj1-soland it will respect your workspace structure. - Save each script in the respective
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.
Identical first prompt
- gpt-4-1106-previewCompare side by side →
- gpt-4-1106-previewCompare side by side →
- gpt-4-1106-previewCompare side by side →
- gpt-4-1106-previewCompare side by side →
- gpt-4-1106-previewCompare side by side →
- gpt-4-1106-previewCompare side by side →
- gpt-4-1106-previewCompare side by side →
- gpt-4-1106-previewCompare side by side →
- gpt-4-1106-previewCompare side by side →
- gpt-4-1106-previewCompare side by side →
- gpt-4-1106-previewCompare side by side →