chemql-python 1.2.1__py3-none-any.whl → 1.2.3__py3-none-any.whl

chemql/__init__.py

@@ -18,6 +18,8 @@ from prompt_toolkit.lexers import Lexer
 import atexit
 import uuid
 import select
+import queue
+import threading
 from prompt_toolkit.styles import Style
 from prompt_toolkit.completion import Completer, Completion
 import urllib.parse
@@ -45,7 +47,7 @@ __all__ = [
     "__version__"
 ]
 
-__version__ = "1.2.1"
+__version__ = "1.2.3"
 
 # Persistent preprocessor/function state for the current ChemQL session.
 DEFINITIONS = {}
@@ -53,6 +55,7 @@ FUNCTIONS = {}
 MACROS = {}
 EXECUTION_HISTORY = []
 _DEFINITION_NAME = r"[A-Za-z_][A-Za-z0-9_]*"
+_NO_RETURN = object()
 
 def _expand_definitions(line):
     """Replace @NAME literally, including inside strings, using current definitions."""
@@ -107,111 +110,203 @@ def _split_function_arguments(source):
     return args
 
 def _call_function(name, arguments):
-    """Call a ChemQL function, including `let` assignments and block syntax."""
+    """Call a ChemQL function, including let assignments and block syntax."""
     if name not in FUNCTIONS:
         raise NameError(f"Unknown ChemQL function `{name}`")
 
     values = [evaluate_python(arg) for arg in arguments]
     definition = FUNCTIONS[name]
+
     saved_globals = python_globals.copy()
     saved_chemql_variables = CHEMQL_VARIABLES.copy()
+
     had_arguments = "arguments" in python_globals
     saved_arguments = python_globals.get("arguments")
-    # `arguments` contains only the supplied arguments, not the function name.
+
+    # Expose only the supplied arguments.
     python_globals["arguments"] = list(values)
 
     def substitute_parameters(source):
-        # Replace positional placeholders as tokens, so \1 does not corrupt \10.
+        """Replace \\1, \\2, etc. without corrupting \\10 or longer indices."""
         return re.sub(
             r"\\(\d+)",
-            lambda match: repr(values[int(match.group(1)) - 1])
-            if 0 < int(match.group(1)) <= len(values) else match.group(0),
+            lambda match: (
+                repr(values[int(match.group(1)) - 1])
+                if 0 < int(match.group(1)) <= len(values)
+                else match.group(0)
+            ),
             source,
         )
 
     def normalize_let_line(source):
-        """Translate `let name = expression` to a Python assignment block."""
-        match = re.fullmatch(r"\s*let\s+([A-Za-z_][A-Za-z0-9_]*)\s*=\s*(.*?)\s*", source)
+        """Translate let name = expression into a Python assignment."""
+        match = re.fullmatch(
+            r"\s*let\s+([A-Za-z_]\w*)\s*=\s*(.*?)\s*",
+            source,
+        )
+
         if not match:
             return source
+
         variable, expression = match.groups()
         expression = substitute_parameters(expression)
-        # In ChemQL function expressions, [i] means evaluate i, rather than
-        # create a Python list containing i. Keep this deliberately simple and
-        # avoid changing arbitrary Python list literals.
+        expression = _expand_definitions(expression)
+
+        # Translate simple ChemQL [variable] expressions into (variable).
         expression = re.sub(
-            r"\[\s*([A-Za-z_][A-Za-z0-9_]*(?:\.[A-Za-z_][A-Za-z0-9_]*)*)\s*\]",
+            r"\[\s*([A-Za-z_]\w*(?:\.[A-Za-z_]\w*)*)\s*\]",
             r"(\1)",
             expression,
         )
+
         return f"[{variable} = {expression}]"
 
-    try:
-        body = definition["body"]
+    def execute_function_lines(lines):
+        """Execute function statements and propagate return values."""
         i = 0
-        while i < len(body):
-            line = body[i].strip()
+
+        while i < len(lines):
+            line = lines[i].strip()
+
             if not line:
                 i += 1
                 continue
 
-            # A function may contain a multiline [[[ Python block ]]].
+            # Multiline Python block.
             if line == "[[[":
                 block = []
                 i += 1
-                while i < len(body) and body[i].strip() != "]]]":
-                    block.append(substitute_parameters(body[i]))
+
+                while i < len(lines) and lines[i].strip() != "]]]":
+                    source_line = substitute_parameters(lines[i])
+                    source_line = _expand_definitions(source_line)
+                    block.append(source_line)
                     i += 1
-                if i >= len(body):
-                    raise SyntaxError(f"Missing `]]]` in function `{name}`")
+
+                if i >= len(lines):
+                    raise SyntaxError(
+                        f"Missing `]]]` in function `{name}`"
+                    )
+
                 evaluate_python("\n".join(block))
                 i += 1
                 continue
 
-            # Support ChemQL's [for ...], [if ...], and [while ...] blocks
-            # inside functions, using the same block executors as the REPL.
-            if is_python_block_header(line):
-                end = find_endblock(body, i + 1)
+            # Recognize block headers generically, while excluding
+            # closing and branch markers from being treated as headers.
+            generic_header = re.match(
+                r"^\[\s*(?!endblock\b|else\b|elif\b)"
+                r"[A-Za-z_]\w*\b.*\]\s*$",
+                line,
+            )
+
+            if is_python_block_header(line) or generic_header:
+                end = find_endblock(lines, i + 1)
+
                 block_body = [
                     normalize_let_line(substitute_parameters(source_line))
-                    for source_line in body[i + 1:end]
+                    for source_line in lines[i + 1:end]
                 ]
+
                 code = line[1:-1].strip()
+
                 if code.startswith("for "):
                     execute_for_block(line, block_body)
+
                 elif code.startswith("while "):
                     execute_while_block(line, block_body)
+
                 elif code.startswith("if "):
-                    execute_if_block(line, block_body)
+                    # Expand parameters and definitions before evaluating.
+                    condition = code[3:].strip()
+                    condition = substitute_parameters(condition)
+                    condition = _expand_definitions(condition)
+
+                    # Find the top-level else branch.
+                    else_index = next(
+                        (
+                            j
+                            for j, source_line in enumerate(block_body)
+                            if source_line.strip() == "[else]"
+                        ),
+                        None,
+                    )
+
+                    if evaluate_python(condition):
+                        selected_body = (
+                            block_body
+                            if else_index is None
+                            else block_body[:else_index]
+                        )
+                    else:
+                        selected_body = (
+                            []
+                            if else_index is None
+                            else block_body[else_index + 1:]
+                        )
+
+                    # Execute only the selected branch. A return exits
+                    # this function immediately.
+                    result = execute_function_lines(selected_body)
+
+                    if result is not _NO_RETURN:
+                        return result
+
                 else:
-                    raise SyntaxError(f"Unsupported function block `{code}`")
+                    raise SyntaxError(
+                        f"Unsupported function block `{code}`"
+                    )
+
                 i = end + 1
                 continue
 
             if line == "[endblock]":
-                raise SyntaxError(f"Unexpected `[endblock]` in function `{name}`")
+                raise SyntaxError(
+                    f"Unexpected `[endblock]` in function `{name}`"
+                )
+
+            # Handle return statements inside the function.
+            if re.match(r"^return\b", line):
+                expression = re.sub(r"^return\b", "", line, count=1).strip()
 
-            if line.startswith("return"):
-                expression = line[len("return"):].strip()
                 if expression.startswith("[") and expression.endswith("]"):
-                    expression = expression[1:-1]
+                    expression = expression[1:-1].strip()
+
                 expression = substitute_parameters(expression)
+                expression = _expand_definitions(expression)
+
                 return evaluate_python(expression) if expression else None
 
             normalized = normalize_let_line(line)
+
             if normalized != line:
-                # The wrapper is ChemQL syntax; pass the assignment itself to Python.
                 evaluate_python(normalized[1:-1])
+            elif line.startswith("[") and line.endswith("]"):
+                expression = line[1:-1].strip()
+                expression = substitute_parameters(expression)
+                expression = _expand_definitions(expression)
+                evaluate_python(expression)
             else:
-                evaluate_python(substitute_parameters(line))
+                expression = substitute_parameters(line)
+                expression = _expand_definitions(expression)
+                evaluate_python(expression)
+
             i += 1
-        return None
+
+        return _NO_RETURN
+
+    try:
+        result = execute_function_lines(definition["body"])
+        return None if result is _NO_RETURN else result
+
     finally:
-        # Keep function-local assignments from leaking into the caller's namespace.
+        # Restore the caller's namespaces.
         python_globals.clear()
         python_globals.update(saved_globals)
+
         CHEMQL_VARIABLES.clear()
         CHEMQL_VARIABLES.update(saved_chemql_variables)
+
         if had_arguments:
             python_globals["arguments"] = saved_arguments
         else:
@@ -258,17 +353,15 @@ _shell_process = None
 _shell_master_fd = None
 _shell_prompt_ready = False
 _shell_output_streamed = False
+_shell_output_queue = None
+_shell_reader_thread = None
 _shell_prompt = f"__CHEMQL_PROMPT_{uuid.uuid4().hex}__"
 
 
 def get_shell():
-    """Return the persistent shell process.
-
-    On POSIX, use a pseudo-terminal so foreground commands such as `read`,
-    `read -p`, and interactive CLIs can receive keyboard input. Windows keeps
-    the pipe-based implementation until a ConPTY backend is added.
-    """
+    """Return the persistent OS shell process for the current platform."""
     global _shell_process, _shell_master_fd, _shell_prompt_ready
+    global _shell_output_queue, _shell_reader_thread
 
     if _shell_process is not None and _shell_process.poll() is None:
         return _shell_process
@@ -277,14 +370,38 @@ def get_shell():
     close_shell()
 
     if os.name == "nt":
+        # Use PowerShell as ChemQL's native Windows shell. Its prompt function
+        # emits a private marker when a command finishes, so we do not append a
+        # marker to stdin (which would accidentally become Read-Host input).
+        setup = (
+            f"function global:prompt {{ [Console]::Out.WriteLine('{_shell_prompt}'); return '' }}"
+        )
         _shell_process = subprocess.Popen(
-            ["cmd.exe", "/D", "/Q", "/K", "prompt $S"],
+            ["powershell.exe", "-NoLogo", "-NoProfile", "-NoExit", "-Command", setup],
             stdin=subprocess.PIPE,
             stdout=subprocess.PIPE,
             stderr=subprocess.STDOUT,
             text=True,
             bufsize=1,
         )
+        output_queue = queue.Queue()
+        _shell_output_queue = output_queue
+        process = _shell_process
+
+        def read_output():
+            try:
+                while True:
+                    character = process.stdout.read(1)  # type: ignore
+                    if not character:
+                        break
+                    output_queue.put(character)
+            except (OSError, ValueError):
+                pass
+            finally:
+                output_queue.put(None)
+
+        _shell_reader_thread = threading.Thread(target=read_output, daemon=True)
+        _shell_reader_thread.start()
     else:
         import fcntl
         import pty
@@ -434,47 +551,128 @@ def _run_shell_command_posix(command, marker=None):
         return f"Shell error: {exc}"
 
 
-def run_shell_command(command):
+def _run_shell_command_windows(command):
+    """Run a command in persistent PowerShell and forward line input interactively."""
+    global _shell_prompt_ready, _shell_output_streamed
     shell = get_shell()
-    command = command.replace("\r\n", "\n").replace("\r", "\n").rstrip("\n")
-    marker = f"__CHEMQL_END_{uuid.uuid4().hex}__"
+    output_queue = _shell_output_queue
+    if output_queue is None:
+        return "Shell error: PowerShell output reader was not initialized."
 
-    try:
-        if os.name != "nt":
-            return _run_shell_command_posix(command, marker)
+    import msvcrt
+
+    def wait_for_prompt(stream_output):
+        collected = ""
+        streamed_pos = 0
+        marker = _shell_prompt
+        marker_len = len(marker)
+        stdin_is_tty = False
+        try:
+            stdin_is_tty = sys.stdin is not None and sys.stdin.isatty()
+        except (AttributeError, OSError, ValueError):
+            pass
+        pending_input = ""
 
-        # Windows pipe mode supports persistent commands, but not fully
-        # interactive programs yet; a ConPTY backend is needed for that.
-        shell.stdin.write(command + "\r\n" + f"echo {marker}\r\n")  # type: ignore
-        shell.stdin.flush()  # type: ignore
-        output = []
         while True:
+            # While a shell command is running, pass complete lines typed by the
+            # user to PowerShell. This lets Read-Host receive real input rather
+            # than consuming a completion marker injected by ChemQL.
+            if stdin_is_tty:
+                try:
+                    while msvcrt.kbhit(): #type:ignore
+                        key = msvcrt.getwch()  #type:ignore
+                        if key in ("\x00", "\xe0"):  #type:ignore
+                            # Consume the second half of a Windows special key.
+                            if msvcrt.kbhit():  #type:ignore
+                                msvcrt.getwch()  #type:ignore
+                        elif key in ("\r", "\n"):
+                            shell.stdin.write(pending_input + "\r\n")  # type: ignore
+                            shell.stdin.flush()  # type: ignore
+                            pending_input = ""
+                        elif key == "\b":
+                            pending_input = pending_input[:-1]
+                        elif key == "\x03":
+                            # Ctrl-C is sent as a line input; reset is safer than
+                            # leaving ChemQL stuck in a nested interactive prompt.
+                            close_shell()
+                            return None
+                        elif key >= " " and key != "\x7f":
+                            pending_input += key
+                except (OSError, ValueError):
+                    stdin_is_tty = False
+
             try:
-                line = shell.stdout.readline()  # type: ignore
-            except KeyboardInterrupt:
-                # Avoid leaking a traceback; reset the shell so the next command
-                # starts with a clean process if the current one was interrupted.
-                close_shell()
-                return "Shell command interrupted. Shell session was reset."
-            if not line:
-                return "\n".join(output) or "Shell session terminated."
-            line = line.rstrip("\r\n")
-            if marker in line:
-                break
-            output.append(line)
-        return "\n".join(output).strip()
+                character = output_queue.get(timeout=0.04)
+            except queue.Empty:
+                continue
+            if character is None:
+                return None
 
+            collected += character
+            marker_pos = collected.find(marker)
+            if marker_pos >= 0:
+                visible = collected[:marker_pos]
+                if stream_output and marker_pos > streamed_pos:
+                    sys.stdout.write(collected[streamed_pos:marker_pos])
+                    sys.stdout.flush()
+                    _shell_output_streamed = True
+                return visible.replace("\r\n", "\n").rstrip("\r\n ")
+
+            if stream_output:
+                # Stream immediately, retaining the full output for history and
+                # only withholding a suffix that could begin the prompt marker.
+                overlap = 0
+                max_overlap = min(marker_len - 1, len(collected))
+                for size in range(max_overlap, 0, -1):
+                    if collected.endswith(marker[:size]):
+                        overlap = size
+                        break
+                safe_end = len(collected) - overlap
+                if safe_end > streamed_pos:
+                    sys.stdout.write(collected[streamed_pos:safe_end])
+                    sys.stdout.flush()
+                    _shell_output_streamed = True
+                    streamed_pos = safe_end
+
+    if not _shell_prompt_ready:
+        initial = wait_for_prompt(stream_output=False)
+        if initial is None:
+            close_shell()
+            return "Shell session terminated."
+        _shell_prompt_ready = True
+
+    try:
+        shell.stdin.write(command.replace("\n", "\r\n") + "\r\n")  # type: ignore
+        shell.stdin.flush()  # type: ignore
+        result = wait_for_prompt(stream_output=True)
+        if result is None:
+            close_shell()
+            return "Shell session terminated."
+        return result
+    except (BrokenPipeError, OSError, ValueError) as exc:
+        close_shell()
+        return f"Shell error: {exc}"
+
+
+def run_shell_command(command):
+    global _shell_output_streamed
+    command = command.replace("\r\n", "\n").replace("\r", "\n").rstrip("\n")
+    try:
+        if os.name == "nt":
+            if command.strip().casefold() in {"powershell", "powershell.exe", "pwsh", "pwsh.exe"}:
+                return "ChemQL already uses PowerShell on Windows; run the command directly instead of starting a nested shell."
+            _shell_output_streamed = False
+            return _run_shell_command_windows(command)
+        return _run_shell_command_posix(command)
     except KeyboardInterrupt:
         close_shell()
         return "Shell command interrupted. Shell session was reset."
-    except (BrokenPipeError, OSError) as exc:
-        close_shell()
-        return f"Shell error: {exc}"
 
 
 def close_shell():
     """Clean up the persistent shell and its pseudo-terminal on exit."""
     global _shell_process, _shell_master_fd, _shell_prompt_ready
+    global _shell_output_queue, _shell_reader_thread
 
     if _shell_process is not None:
         if _shell_process.poll() is None:
@@ -496,6 +694,8 @@ def close_shell():
             pass
         _shell_master_fd = None
     _shell_prompt_ready = False
+    _shell_output_queue = None
+    _shell_reader_thread = None
 
 
 atexit.register(close_shell)
@@ -2486,7 +2686,7 @@ def execute_if_block(header, body):
     )
 
     for section_header, section_body in sections:
-        code = section_header[1:-1].strip()
+        code = _expand_definitions(section_header[1:-1].strip())
 
         if code.startswith("if "):
             condition = code[3:].strip()
@@ -3229,7 +3429,7 @@ def execute(arguments, query_text=None, positions=None):
         
     if arguments[0].lower() == "macros":
         return "\n".join([f"@{name} = {value}" for name, value in MACROS.items()]) if MACROS else "No macros."
-
+    
     if arguments[0] == "conditions":
         if len(arguments) != 1:
             return "Usage: conditions"
@@ -3584,7 +3784,7 @@ class ChemqlLexer(Lexer):
         "findmol", "findre", "set", "add", "remove", "react",
         "balance", "conditions", "view", "help", "dump", "version",
         "return", "union", "intersection", "let", "unset", "variables",
-        "macros", "definitions", "arguments"
+        "macros", "definitions", "arguments",
     }
 
     OPERATORS = {

{chemql_python-1.2.1.dist-info → chemql_python-1.2.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: chemql-python
-Version: 1.2.1
+Version: 1.2.3
 Summary: A chemistry-focused query language and REPL
 Author-email: Mohammad Rakibul Islam <rakibulofficial1017@gmail.com>
 Requires-Python: >=3.10
@@ -17,15 +17,15 @@ Dynamic: license-file
 shell commands in one interactive environment.**
 
 ChemQL is a command-line chemistry query language for exploring chemical
-elements, molecules, and reactions. Version 1.2.0 combines database-style
+elements, molecules, and reactions. Version 1.2.3 combines database-style
 chemistry queries, reaction tools, Python interoperability, reusable functions
 and macros, and a persistent shell in one scriptable environment. It combines a small SQL-inspired
 query syntax with a persistent Python environment, reaction utilities,
 optional external molecule lookup, and tools for running repeatable
 scripts.
 
-> **Version covered by this README:** 1.2.0\
-> The examples below describe the ChemQL 1.2.0 source. Dataset
+> **Version covered by this README:** 1.2.3\
+> The examples below describe the ChemQL 1.2.3 source. Dataset
 > contents and external-service availability can affect individual
 > lookup results.
 
@@ -538,7 +538,7 @@ queried or inspected from Python.
 
 ## Preprocessor, functions, and macros
 
-ChemQL 1.2.0 adds three ways to reuse values and logic. These declarations
+ChemQL 1.2.3 adds three ways to reuse values and logic. These declarations
 are processed in source order: definitions affect later lines, not earlier
 ones. They persist during the current ChemQL process/session; they are not
 saved automatically between separate launches.
@@ -833,7 +833,7 @@ redirect a file into the command.
 ``` bash
 chemql --version
 chemql --upgrade
-chemql --upgrade 1.2.0
+chemql --upgrade 1.2.3
 ```
 
 `--upgrade` installs the latest published `chemql-python` release; supplying a

{chemql_python-1.2.1.dist-info → chemql_python-1.2.3.dist-info}/RECORD

@@ -1,4 +1,4 @@
-chemql/__init__.py,sha256=7YotC6BZV7ZRPQjQwFfmOmlbri9tdKPloQ40Lm_nRZg,152580
+chemql/__init__.py,sha256=buPnjHRRx32NB8I6hXUgKoKLzy3R_FkdNYqm1mmptLA,159524
 chemql/__main__.py,sha256=3Q52s0hyuouhvSSqFEQFDr6mtJ-azkCwI2AD4c7f7A8,28
 chemql/data/BondEnergies.json,sha256=QZN3iWZrD8125jV41EG1GiRohBtP0CzAP_EDQTfaEKc,8269
 chemql/data/MoleculeBonds.json,sha256=Iw2shYxJScH_tI80IytBYoqVDSbgT7OSqPJUIgKZme4,84262
@@ -8,9 +8,9 @@ chemql/data/Reactions.json,sha256=BFpkKEWwLzG54ot96BVOWwU3btpimRJlOM57yGclGgc,15
 chemql/data/THIRD_PARTY_LICENSES/PeriodicTableJSON.txt,sha256=9wih5q8s58qR3vJEGpooa4ESyAGZHcmGvzwMMkMHNKk,322
 chemql/data/THIRD_PARTY_LICENSES/RHEA.txt,sha256=nXTeCP6vJBW5B_mtaVGFLr0w7vnjrSTBnv6_eOk3KKo,634
 chemql/html/periodic_table.html,sha256=ftVgXWjRO2AsTAGEekTrvgE5ZvcwRfoby9UWoixc9Y8,361560
-chemql_python-1.2.1.dist-info/licenses/LICENSE,sha256=o4ywHC0EwRy2zgIASiRD2sK3Vcsf7LBcTHjDvoJgRE0,1079
-chemql_python-1.2.1.dist-info/METADATA,sha256=iKKC3niSxAHMvcc0CxxgfeDw9EhQDx0RZXt7_kdbDck,30529
-chemql_python-1.2.1.dist-info/WHEEL,sha256=YVMoNqKzERt-wjUZwJ33xBGAwnFl-4cqbYkTtWa4itE,91
-chemql_python-1.2.1.dist-info/entry_points.txt,sha256=j98ObVWR743an524z24AiWeRsppag-py5cc7l0D0HkI,39
-chemql_python-1.2.1.dist-info/top_level.txt,sha256=9UzHwnk_F6W19Q9nayaFgVyAELQXXwoYjeBgB6MWNr8,7
-chemql_python-1.2.1.dist-info/RECORD,,
+chemql_python-1.2.3.dist-info/licenses/LICENSE,sha256=o4ywHC0EwRy2zgIASiRD2sK3Vcsf7LBcTHjDvoJgRE0,1079
+chemql_python-1.2.3.dist-info/METADATA,sha256=HMq0gKLpGQrzdy9Ycj5evg8E0D57_Hfr28TSNOQHuXM,30529
+chemql_python-1.2.3.dist-info/WHEEL,sha256=YVMoNqKzERt-wjUZwJ33xBGAwnFl-4cqbYkTtWa4itE,91
+chemql_python-1.2.3.dist-info/entry_points.txt,sha256=j98ObVWR743an524z24AiWeRsppag-py5cc7l0D0HkI,39
+chemql_python-1.2.3.dist-info/top_level.txt,sha256=9UzHwnk_F6W19Q9nayaFgVyAELQXXwoYjeBgB6MWNr8,7
+chemql_python-1.2.3.dist-info/RECORD,,