Merge pull request #19 from iwyoo/user-given-temp-filepath

feat: Add temporary-file-path argument for multi-processing

Author: JacksonBurns

README.md

@@ -38,10 +38,11 @@ smiles_str = "CC"
     allow_bad_stereo = False,
     wildcard_radicals = False,
     jar_fpath = "/path/to/opsin.jar",
+    tmp_fpath = "py2opsin_temp_input.txt",
 )
 ```
 
-The result is returned as a Python string, or False if an unexpected error occurs when calling OPSIN. If a list of IUPAC names is provided, a list is returned. It is __highly__ reccomended to use `py2opsin` in this manner if you need to resolve any more than a couple names -- the performance cost of running `OPSIN` from Python one name at a time is significant (~5 seconds/molecule individually, milliseconds otherwise).
+The result is returned as a Python string, or False if an unexpected error occurs when calling OPSIN. If a list of IUPAC names is provided, a list is returned. It is __highly__ recommended to use `py2opsin` in this manner if you need to resolve any more than a couple names -- the performance cost of running `OPSIN` from Python one name at a time is significant (~5 seconds/molecule individually, milliseconds otherwise).
 
 Arguments:
  - chemical_name (str): IUPAC name of chemical as a Python string, or a list of strings.
@@ -51,7 +52,10 @@ Arguments:
  - allow_bad_stereo (bool, optional): Allow OPSIN to ignore uninterpreatable stereochem. Defaults to False.
  - wildcard_radicals (bool, optional): Output radicals as wildcards. Defaults to False.
  - jar_fpath (str, optional): Filepath to OPSIN jar file. Defaults to "opsin-cli.jar" which is distributed with py2opsin.
+ - tmp_fpath (str, optional): tmp_fpath (str, optional): Name for temporary file used for calling OPSIN. Defaults to "py2opsin_temp_input.txt". When multiprocessing, set this to a unique name for each process.
 
+> [!TIP]
+> `OPSIN` will already parallelize itself by creating multiple threads! Be wary when using `py2opsin` with multiprocessing to avoid spawning too many processes.
 
 ## Massive speedup from `pubchempy` for batch translations
 `py2opsin` runs locally and is smaller in scope in what it provides, which makes it __dramatically__ faster at resolving identifiers. In the code block below, the call to `py2opsin` will execute faster than an equivalent call to `pubchempy`:

py2opsin/__init__.py

@@ -1,3 +1,3 @@
 from .py2opsin import py2opsin
 
-__version__ = "1.0.6"
+__version__ = "1.1.0"

py2opsin/py2opsin.py

@@ -53,6 +53,7 @@ def py2opsin(
     allow_bad_stereo: bool = False,
     wildcard_radicals: bool = False,
     jar_fpath: str = "default",
+    tmp_fpath: str = "py2opsin_temp_input.txt",
 ) -> str:
     """Simple passthrough to opsin, returning results as Python strings.
 
@@ -65,6 +66,8 @@ def py2opsin(
         allow_bad_stereo (bool, optional): Allow OPSIN to ignore uninterpreatable stereochem. Defaults to False.
         wildcard_radicals (bool, optional): Output radicals as wildcards. Defaults to False.
         jar_fpath (str, optional): Filepath to OPSIN jar file. Defaults to "default", which causes py2opsin to use its included jar.
+        tmp_fpath (str, optional): Name for temporary file used for calling OPSIN. Defaults to "py2opsin_temp_input.txt".
+                                   When multiprocessing, set this to a unique name for each process.
 
     Returns:
         str: Species in requested format, or False if not found or an error ocurred. List of strings if input is list.
@@ -112,15 +115,14 @@ def py2opsin(
         )
 
     # write the input to a text file
-    temp_f = "py2opsin_temp_input.txt"
-    with open(temp_f, "w") as file:
+    with open(tmp_fpath, "w") as file:
         if type(chemical_name) is str:
             file.write(chemical_name)
         else:
             file.writelines("\n".join(chemical_name) + "\n")
 
     # add the temporary file to the args
-    arg_list.append(temp_f)
+    arg_list.append(tmp_fpath)
 
     # grab the optional boolean flags
     if allow_acid:
@@ -168,4 +170,4 @@ def py2opsin(
         warnings.warn("Unexpected error ocurred! " + e)
         return False
     finally:
-        os.remove(temp_f)
+        os.remove(tmp_fpath)

test/test_py2opsin.py

@@ -1,10 +1,16 @@
 import os
+import multiprocessing
 import sys
 import unittest
 
 from py2opsin import py2opsin
 
 
+# multiprocessing test function
+def _f(b):
+    return py2opsin(b[0], tmp_fpath=f"tmp_{b[1]}.txt")
+
+
 class Test_py2opsin(unittest.TestCase):
     """
     Test the various functionalities of py2opsin.
@@ -97,6 +103,12 @@ def test_invalid_output_helpful_error(self):
             "Output format SMOLES is invalid. Did you mean 'SMILES'?",
         )
 
+    def test_multiprocessing(self):
+        """py2opsin should safely work when run with multiprocessing"""
+        with multiprocessing.Pool(2) as pool:
+            res = pool.map(_f, [("methanol", 0), ("ethanol", 1)])
+        self.assertEqual(res, ["CO", "C(C)O"])
+
     def test_name_to_smiles(self):
         """
         Tests converting IUPAC names to SMILES strings