psdi-data-conversion 0.0.38__tar.gz → 0.0.39__tar.gz

{psdi_data_conversion-0.0.38 → psdi_data_conversion-0.0.39}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: psdi_data_conversion
-Version: 0.0.38
+Version: 0.0.39
 Summary: Chemistry file format conversion service, provided by PSDI
 Project-URL: Homepage, https://data-conversion.psdi.ac.uk/
 Project-URL: Documentation, https://psdi-uk.github.io/psdi-data-conversion/

{psdi_data_conversion-0.0.38 → psdi_data_conversion-0.0.39}/psdi_data_conversion/app.py

@@ -5,22 +5,28 @@ Version 1.0, 8th November 2024
 This script acts as a server for the PSDI Data Conversion Service website.
 """
 
-from argparse import ArgumentParser
-import hashlib
-import os
 import json
+import os
+import sys
+from argparse import ArgumentParser
+from collections.abc import Callable
 from datetime import datetime
+from functools import wraps
+from hashlib import md5
 from subprocess import run
-import sys
-import traceback
-from flask import Flask, request, render_template, abort, Response
+from traceback import format_exc
+from typing import Any
+
+import werkzeug.serving
+from flask import Flask, Response, abort, cli, render_template, request
 
 import psdi_data_conversion
-from psdi_data_conversion import log_utility
 from psdi_data_conversion import constants as const
+from psdi_data_conversion import log_utility
 from psdi_data_conversion.converter import run_converter
-from psdi_data_conversion.database import FormatInfo, get_format_info
+from psdi_data_conversion.database import get_format_info
 from psdi_data_conversion.file_io import split_archive_ext
+from psdi_data_conversion.main import print_wrap
 
 # Env var for the SHA of the latest commit
 SHA_EV = "SHA"
@@ -36,7 +42,7 @@ FILE_TO_UPLOAD_KEY = 'fileToUpload'
 
 # Create a token by hashing the current date and time.
 dt = str(datetime.now())
-token = hashlib.md5(dt.encode('utf8')).hexdigest()
+token = md5(dt.encode('utf8')).hexdigest()
 
 # Get the service and production modes from their envvars
 service_mode_ev = os.environ.get(SERVICE_MODE_EV)
@@ -80,6 +86,22 @@ if ev_max_file_size_ob is not None:
 else:
     max_file_size_ob = const.DEFAULT_MAX_FILE_SIZE_OB
 
+# Since we're using the development server as the user GUI, we monkey-patch Flask to disable the warnings that would
+# otherwise appear for this so they don't confuse the user
+
+
+def suppress_warning(func: Callable[..., Any]) -> Callable[..., Any]:
+    @wraps(func)
+    def wrapper(*args, **kwargs) -> Any:
+        if args and isinstance(args[0], str) and args[0].startswith('WARNING: This is a development server.'):
+            return ''
+        return func(*args, **kwargs)
+    return wrapper
+
+
+werkzeug.serving._ansi_style = suppress_warning(werkzeug.serving._ansi_style)
+cli.show_server_banner = lambda *_: None
+
 app = Flask(__name__)
 
 
@@ -101,7 +123,7 @@ def get_last_sha() -> str:
         out_str = str(out_bytes.decode()).strip()
 
     except Exception:
-        print("ERROR: Could not determine SHA of most recent commit. Error was:\n" + traceback.format_exc(),
+        print("ERROR: Could not determine SHA of most recent commit. Error was:\n" + format_exc(),
               file=sys.stderr)
         out_str = "N/A"
 
@@ -146,7 +168,7 @@ def convert():
         name = request.form[format_label]
         full_note = request.form[format_label+"_full"]
 
-        l_possible_formats: list[FormatInfo] = get_format_info(name, which="all")
+        l_possible_formats = get_format_info(name, which="all")
 
         # If there's only one possible format, use that
         if len(l_possible_formats) == 1:
@@ -154,13 +176,11 @@ def convert():
             continue
 
         # Otherwise, find the format with the matching note
-        found = False
         for possible_format in l_possible_formats:
             if possible_format.note in full_note:
                 d_formats[format_label] = possible_format
-                found = True
                 break
-        if not found:
+        else:
             print(f"Format '{name}' with full description '{full_note}' could not be found in database.",
                   file=sys.stderr)
             abort(const.STATUS_CODE_GENERAL)
@@ -200,7 +220,7 @@ def convert():
             else:
                 # Failsafe exception message
                 msg = ("The following unexpected exception was raised by the converter:\n" +
-                       traceback.format_exc()+"\n")
+                       format_exc()+"\n")
             with open(qualified_output_log, "w") as fo:
                 fo.write(msg)
             abort(status_code)
@@ -360,6 +380,10 @@ def main():
         global log_level
         log_level = args.log_level
 
+    print_wrap("Starting the PSDI Data Conversion GUI. This GUI is run as a webpage, which you can open by "
+               "right-clicking the link below to open it in your default browser, or by copy-and-pasting it into your "
+               "browser of choice.")
+
     start_app()
 
 

{psdi_data_conversion-0.0.38 → psdi_data_conversion-0.0.39}/psdi_data_conversion/converter.py

@@ -5,18 +5,18 @@ Created 2024-12-10 by Bryan Gillis.
 Class and functions to perform file conversion
 """
 
-from dataclasses import dataclass, field
-import os
+import glob
 import importlib
+import os
 import sys
-from tempfile import TemporaryDirectory
 import traceback
-from typing import Any, Callable, NamedTuple
+from collections.abc import Callable
+from dataclasses import dataclass, field
+from tempfile import TemporaryDirectory
+from typing import Any, NamedTuple
+
 from psdi_data_conversion import constants as const
 from psdi_data_conversion.converters import base
-
-import glob
-
 from psdi_data_conversion.converters.openbabel import CONVERTER_OB
 from psdi_data_conversion.file_io import (is_archive, is_supported_archive, pack_zip_or_tar, split_archive_ext,
                                           unpack_zip_or_tar)

{psdi_data_conversion-0.0.38 → psdi_data_conversion-0.0.39}/psdi_data_conversion/converters/base.py

@@ -6,19 +6,19 @@ Base class and information for file format converters
 """
 
 
-from copy import deepcopy
-from dataclasses import dataclass
+import abc
 import logging
-from collections.abc import Callable
 import os
 import subprocess
-import abc
-
 import sys
 import traceback
+from collections.abc import Callable
+from copy import deepcopy
+from dataclasses import dataclass
 from typing import Any
 
-from psdi_data_conversion import constants as const, log_utility
+from psdi_data_conversion import constants as const
+from psdi_data_conversion import log_utility
 from psdi_data_conversion.dist import bin_exists, get_bin_path, get_dist
 from psdi_data_conversion.security import SAFE_STRING_RE, string_is_safe
 
@@ -735,24 +735,11 @@ class ScriptFileConverter(FileConverter):
 
         self.logger.debug(f"Performing conversion with ScriptFileConverter using script '{self.script}'")
 
-        from_flags = self.data.get("from_flags", "")
-        to_flags = self.data.get("from_flags", "")
-        from_options = self.data.get("from_options", "")
-        to_options = self.data.get("from_options", "")
-
-        # Check that all user-provided input passes security checks
-        for user_args in [from_flags, to_flags, from_options, to_options]:
-            if not string_is_safe(user_args):
-                raise FileConverterInputException(f"Provided argument '{user_args}' does not pass security check - it "
-                                                  f"must match the regex {SAFE_STRING_RE.pattern}.", help=True)
-
         env = {"DIST": get_dist()}
         if self.required_bin is not None:
             env["BIN_PATH"] = get_bin_path(self.required_bin)
 
-        process = subprocess.run(['sh', f'psdi_data_conversion/scripts/{self.script}',
-                                  '--' + self.to_format_info.name, self.in_filename, self.out_filename, from_flags,
-                                  to_flags, from_options, to_options],
+        process = subprocess.run(['sh', f'psdi_data_conversion/scripts/{self.script}', *self._get_script_args()],
                                  env=env, capture_output=True, text=True)
 
         self.out = process.stdout
@@ -763,3 +750,20 @@ class ScriptFileConverter(FileConverter):
             self._abort_from_err()
         else:
             self.logger.debug("Conversion process completed successfully")
+
+    def _get_script_args(self):
+        """Get the list of arguments which will be passed to the script"""
+
+        from_flags = self.data.get("from_flags", "")
+        to_flags = self.data.get("from_flags", "")
+        from_options = self.data.get("from_options", "")
+        to_options = self.data.get("from_options", "")
+
+        # Check that all user-provided input passes security checks
+        for user_args in [from_flags, to_flags, from_options, to_options]:
+            if not string_is_safe(user_args):
+                raise FileConverterInputException(f"Provided argument '{user_args}' does not pass security check - it "
+                                                  f"must match the regex {SAFE_STRING_RE.pattern}.", help=True)
+
+        return ['--' + self.to_format_info.name, self.in_filename, self.out_filename, from_flags, to_flags,
+                from_options, to_options]

{psdi_data_conversion-0.0.38 → psdi_data_conversion-0.0.39}/psdi_data_conversion/converters/c2x.py

@@ -29,6 +29,19 @@ class C2xFileConverter(ScriptFileConverter):
             "may be downloaded from https://www.c2x.org.uk/downloads/")
     supports_ambiguous_extensions = True
 
+    def _get_script_args(self):
+        """Override the standard script arguments so we can set the different format names expected by c2x
+        """
+        l_script_args = super()._get_script_args()
+
+        # Update the output format to c2x style
+        l_script_args[0] = "--" + self.to_format_info.c2x_format
+
+        # TODO - check if the input file has an extension which will be accepted by c2x for its format, and handle if
+        # not
+
+        return l_script_args
+
 
 # Assign this converter to the `converter` variable - this lets the psdi_data_conversion.converter module detect and
 # register it, making it available for use by the command-line script, python library, and web app

{psdi_data_conversion-0.0.38 → psdi_data_conversion-0.0.39}/psdi_data_conversion/converters/openbabel.py

@@ -6,8 +6,9 @@ Open Babel FileConverter
 """
 
 from copy import deepcopy
-from openbabel import openbabel
+
 import py
+from openbabel import openbabel
 
 from psdi_data_conversion.converters.base import FileConverter, FileConverterInputException
 from psdi_data_conversion.security import SAFE_STRING_RE, string_is_safe

{psdi_data_conversion-0.0.38 → psdi_data_conversion-0.0.39}/psdi_data_conversion/database.py

@@ -7,12 +7,12 @@ Python module provide utilities for accessing the converter database
 
 from __future__ import annotations
 
+import json
+import os
 from dataclasses import dataclass, field
 from itertools import product
-import json
 from logging import getLogger
-import os
-from typing import Any, Literal
+from typing import Any, Literal, overload
 
 from psdi_data_conversion import constants as const
 from psdi_data_conversion.converter import D_REGISTERED_CONVERTERS, D_SUPPORTED_CONVERTERS
@@ -32,6 +32,7 @@ DB_URL_KEY = "url"
 
 # Keys for format general info in the database
 DB_FORMAT_EXT_KEY = "extension"
+DB_FORMAT_C2X_KEY = "format"
 DB_FORMAT_NOTE_KEY = "note"
 DB_FORMAT_COMP_KEY = "composition"
 DB_FORMAT_CONN_KEY = "connections"
@@ -428,6 +429,9 @@ class FormatInfo:
         self.id: int = d_single_format_info.get(DB_ID_KEY, -1)
         """The ID of this format"""
 
+        self.c2x_format: str = d_single_format_info.get(DB_FORMAT_C2X_KEY)
+        """The name of this format as the c2x converter expects it"""
+
         self.note: str = d_single_format_info.get(DB_FORMAT_NOTE_KEY, "")
         """The description of this format"""
 
@@ -708,8 +712,8 @@ class ConversionsTable:
             conversion, the second is the info of the input format for this conversion, and the third is the info of the
             output format
         """
-        l_in_format_infos: list[FormatInfo] = self.parent.get_format_info(in_format, which="all")
-        l_out_format_infos: list[FormatInfo] = self.parent.get_format_info(out_format, which="all")
+        l_in_format_infos = self.parent.get_format_info(in_format, which="all")
+        l_out_format_infos = self.parent.get_format_info(out_format, which="all")
 
         # Start a list of all possible conversions
         l_possible_conversions = []
@@ -934,6 +938,16 @@ class DataConversionDatabase:
                                                  f" of type '{type(converter_name_or_id)}'. Type must be `str` or "
                                                  "`int`")
 
+    @overload
+    def get_format_info(self,
+                        format_name_or_id: str | int | FormatInfo,
+                        which: int | None = None) -> FormatInfo: ...
+
+    @overload
+    def get_format_info(self,
+                        format_name_or_id: str | int | FormatInfo,
+                        which: Literal["all"]) -> list[FormatInfo]: ...
+
     def get_format_info(self,
                         format_name_or_id: str | int | FormatInfo,
                         which: int | Literal["all"] | None = None) -> FormatInfo | list[FormatInfo]:
@@ -1084,6 +1098,16 @@ def get_converter_info(name: str) -> ConverterInfo:
     return get_database().d_converter_info[name]
 
 
+@overload
+def get_format_info(format_name_or_id: str | int | FormatInfo,
+                    which: int | None = None) -> FormatInfo: ...
+
+
+@overload
+def get_format_info(format_name_or_id: str | int | FormatInfo,
+                    which: Literal["all"]) -> list[FormatInfo]: ...
+
+
 def get_format_info(format_name_or_id: str | int | FormatInfo,
                     which: int | Literal["all"] | None = None) -> FormatInfo | list[FormatInfo]:
     """Gets the information on a given file format stored in the database

{psdi_data_conversion-0.0.38 → psdi_data_conversion-0.0.39}/psdi_data_conversion/dist.py

@@ -7,9 +7,10 @@ Functions and utilities related to handling multiple user OSes and distributions
 
 import os
 import shutil
-import psdi_data_conversion
 import sys
 
+import psdi_data_conversion
+
 # Labels for each platform (which we use for the folder in this project), and the head of the name each platform will
 # have in `sys.platform`
 

{psdi_data_conversion-0.0.38 → psdi_data_conversion-0.0.39}/psdi_data_conversion/file_io.py

@@ -149,8 +149,7 @@ def pack_zip_or_tar(archive_filename: str,
                 archive_format = _format
                 archive_root_filename = split_archive_ext(archive_filename)[0]
                 break
-        # Check that the format was found
-        if archive_format is None:
+        else:
             raise AssertionError("Invalid execution path entered - filename wasn't found with a valid archive "
                                  "extension, but it did pass the `is_supported_archive` check")
     else:

{psdi_data_conversion-0.0.38 → psdi_data_conversion-0.0.39}/psdi_data_conversion/main.py

@@ -7,12 +7,12 @@ Created 2025-01-14 by Bryan Gillis.
 Entry-point file for the command-line interface for data conversion.
 """
 
-from itertools import product
 import logging
-from argparse import ArgumentParser
 import os
 import sys
 import textwrap
+from argparse import ArgumentParser
+from itertools import product
 
 from psdi_data_conversion import constants as const
 from psdi_data_conversion.constants import CL_SCRIPT_NAME, CONVERTER_DEFAULT, TERM_WIDTH
@@ -537,7 +537,7 @@ def detail_format(format_name: str):
     """Prints details on a format
     """
 
-    l_format_info: list[FormatInfo] = get_format_info(format_name, which="all")
+    l_format_info = get_format_info(format_name, which="all")
 
     if len(l_format_info) == 0:
         print_wrap(f"ERROR: Format '{format_name}' not recognised", err=True, newline=True)

{psdi_data_conversion-0.0.38 → psdi_data_conversion-0.0.39}/psdi_data_conversion/static/content/index-versions/psdi-common-footer.html

@@ -17,16 +17,20 @@
       </ul>
       <ul class="footer__col footer__items clean-list">
         <li>
-          <img class="lm-only" src="static/img/ukri-logo-lighttext.png"
-            alt="UKRI logo">
-          <img class="dm-only" src="static/img/ukri-logo-darktext.png"
-            alt="UKRI logo">
+          <a href="https://www.ukri.org/">
+            <img class="lm-only" src="static/img/ukri-logo-lighttext.png"
+              alt="UKRI logo">
+            <img class="dm-only" src="static/img/ukri-logo-darktext.png"
+              alt="UKRI logo">
+          </a>
         </li>
         <li>
-          <img class="lm-only" src="static/img/ukri-epsr-logo-lighttext.png"
-            alt="UKRI EPSR logo">
-          <img class="dm-only" src="static/img/ukri-epsr-logo-darktext.png"
-            alt="UKRI EPSR logo">
+          <a href="https://www.ukri.org/councils/epsrc/">
+            <img class="lm-only" src="static/img/ukri-epsr-logo-lighttext.png"
+              alt="UKRI EPSR logo">
+            <img class="dm-only" src="static/img/ukri-epsr-logo-darktext.png"
+              alt="UKRI EPSR logo">
+          </a>
         </li>
       </ul>
       <div class="footer__hline small-screen-only">

{psdi_data_conversion-0.0.38 → psdi_data_conversion-0.0.39}/psdi_data_conversion/static/content/psdi-common-footer.html

@@ -17,16 +17,20 @@
       </ul>
       <ul class="footer__col footer__items clean-list">
         <li>
-          <img class="lm-only" src="../img/ukri-logo-lighttext.png"
-            alt="UKRI logo">
-          <img class="dm-only" src="../img/ukri-logo-darktext.png"
-            alt="UKRI logo">
+          <a href="https://www.ukri.org/">
+            <img class="lm-only" src="../img/ukri-logo-lighttext.png"
+              alt="UKRI logo">
+            <img class="dm-only" src="../img/ukri-logo-darktext.png"
+              alt="UKRI logo">
+          </a>
         </li>
         <li>
-          <img class="lm-only" src="../img/ukri-epsr-logo-lighttext.png"
-            alt="UKRI EPSR logo">
-          <img class="dm-only" src="../img/ukri-epsr-logo-darktext.png"
-            alt="UKRI EPSR logo">
+          <a href="https://www.ukri.org/councils/epsrc/">
+            <img class="lm-only" src="../img/ukri-epsr-logo-lighttext.png"
+              alt="UKRI EPSR logo">
+            <img class="dm-only" src="../img/ukri-epsr-logo-darktext.png"
+              alt="UKRI EPSR logo">
+          </a>
         </li>
       </ul>
       <div class="footer__hline small-screen-only">