Merge pull request #1 from noah-13/adpate-to-windows

Add Windows compatibility

Author: m-hedderich

.gitignore

@@ -152,6 +152,10 @@ dmypy.json
 # Cython debug symbols
 cython_debug/
 
+# Test file
+test.ipynb
+
+
 # PyCharm
 #  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
 #  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore

documentation/general_example.ipynb

@@ -121,7 +121,7 @@
     "\n",
     "patterns = premise.find_patterns(premise_instances)\n",
     "for p in patterns:\n",
-    "    print(p)"
+    "    print(p)\n"
    ]
   },
   {
@@ -134,9 +134,9 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "pypremise",
+   "display_name": "pypremise (3.13.3)",
    "language": "python",
-   "name": "myenv"
+   "name": "python3"
   },
   "language_info": {
    "codemirror_mode": {
@@ -148,7 +148,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.12.3"
+   "version": "3.13.3"
   }
  },
  "nbformat": 4,

documentation/missclassification_examples.ipynb

@@ -22,7 +22,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
+   "execution_count": 1,
    "id": "e88fc2bd-0e68-4c12-90e4-932e2e550c4c",
    "metadata": {},
    "outputs": [],
@@ -66,7 +66,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 2,
    "id": "020b3ae0",
    "metadata": {},
    "outputs": [
@@ -98,7 +98,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 3,
    "id": "1de56eba-f4d8-4350-83ff-9b6d09a001d9",
    "metadata": {},
    "outputs": [],
@@ -123,7 +123,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": 4,
    "id": "4a63c48e-db46-4247-9044-93aed2302324",
    "metadata": {},
    "outputs": [
@@ -158,7 +158,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 5,
    "id": "5e4d532d-d70c-48e9-bdbb-70cadf7ddf1a",
    "metadata": {},
    "outputs": [],
@@ -203,7 +203,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": 6,
    "id": "6c73c738-3ffc-4b01-b5b7-022a1ff7c652",
    "metadata": {},
    "outputs": [],
@@ -216,7 +216,7 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "pypremise (3.12.3)",
+   "display_name": "pypremise (3.13.3)",
    "language": "python",
    "name": "python3"
   },
@@ -230,7 +230,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.12.3"
+   "version": "3.13.3"
   }
  },
  "nbformat": 4,

pyproject.toml

@@ -19,4 +19,11 @@ packages = ["pypremise"]
 
 
 [tool.setuptools.package-dir]
-"" = "src"
+"" = "src"
+
+[tool.setuptools.package-data]
+pypremise = [
+    "Premise_Applesilicon",
+    "Premise_Linux",
+    "Premise_Windows.exe",
+]

src/pypremise/Premise_Windows.exe

@@ -25,6 +25,7 @@
 import tempfile
 import time
 import logging
+import os
 
 logger = logging.getLogger(__name__)
 
@@ -106,51 +107,90 @@ def __init__(self, voc_index_to_token: Optional[Mapping[int, str]] = None,
         self.group_1_name = group_1_name
         self.premise_engine = premise_engine
 
+
+
+
     def find_patterns(self, instances: List[PremiseInstance]):
         import pypremise.io
-        import os
 
-        # the Premise C++ code reads and write to files for in and output
         feature_file = tempfile.NamedTemporaryFile(delete=False)
+        feature_file.close()
         label_file = tempfile.NamedTemporaryFile(delete=False)
+        label_file.close()
         result_file = tempfile.NamedTemporaryFile(delete=False)
+        result_file.close()
+
+        feature_path = os.path.abspath(feature_file.name).replace("\\", "/")
+        label_path   = os.path.abspath(label_file.name).replace("\\", "/")
+        result_path  = os.path.abspath(result_file.name).replace("\\", "/")
 
-        pypremise.io.write_dat_content(instances, feature_file.name, label_file.name)
+        pypremise.io.write_dat_content(instances, feature_path, label_path)
 
-        # embeddings
+        # === embeddings ===
         if self.embedding_index_to_vector is not None:
             embedding_file = tempfile.NamedTemporaryFile(delete=False)
-            embedding_path = embedding_file.name
+            embedding_file.close()
+            embedding_path = os.path.abspath(embedding_file.name).replace("\\", "/")
+
             max_feature_index = Premise._get_max_feature_index(instances)
-            pypremise.io.write_embedding_file(self.embedding_index_to_vector, embedding_path,
-                                            self.embedding_dimensionality, max_feature_index)
+            pypremise.io.write_embedding_file(
+                self.embedding_index_to_vector,
+                embedding_path,
+                self.embedding_dimensionality,
+                max_feature_index
+            )
         else:
             embedding_file = None
             embedding_path = ""
 
-        # actual Premise
+        # === call Premise C++ program ===
         start_time = time.time()
-        pypremise.io.call_premise_program(feature_file.name, label_file.name, result_file.name, embedding_path,
-                                          self.embedding_dimensionality, self.max_neighbor_distance,
-                                          self.fisher_p_value, self.clause_max_overlap, self.min_overlap,
-                                          self.premise_engine)
-        logger.info(f"Premise ran for {time.time() - start_time} seconds.")
-
-        results = pypremise.io.parse_premise_result(result_file.name, self.group_0_name, self.group_1_name)
-
-        # clean up temporary files
-        os.remove(feature_file.name)
-        os.remove(label_file.name)
-        os.remove(result_file.name)
+        if self.premise_engine is None:
+            self.premise_engine = pypremise.io.get_premise_path()
+        pypremise.io.call_premise_program(
+            feature_path, label_path, result_path, embedding_path,
+            self.embedding_dimensionality, self.max_neighbor_distance,
+            self.fisher_p_value, self.clause_max_overlap, self.min_overlap,
+            self.premise_engine
+        )
+        logger.info(f"Premise ran for {time.time() - start_time:.2f} seconds.")
+
+        # === check result file ===
+        try:
+            size = os.path.getsize(result_path)
+            logger.info(f"Result file size: {size} bytes")
+            if size == 0:
+                logger.warning("Result file is empty — check Premise stderr or parameters.")
+        except Exception as e:
+            logger.error(f"Could not stat result file: {e}")
+
+        # === analyse results ===
+        results = pypremise.io.parse_premise_result(
+            result_path, self.group_0_name, self.group_1_name
+        )
+
+        def safe_remove(path):
+            try:
+                os.remove(path)
+            except PermissionError:
+                time.sleep(0.2)
+                try:
+                    os.remove(path)
+                except Exception:
+                    pass
+
+        for f in [feature_path, label_path, result_path]:
+            safe_remove(f)
         if embedding_file is not None:
-            os.remove(embedding_file.name)
+            safe_remove(embedding_path)
 
-        # if we have a map from indices to tokens, use it to convert our patterns indices to tokens
         if self.voc_index_to_token is not None:
             self._pattern_indices_to_tokens(results)
 
         return results
 
+
+
     def _pattern_indices_to_tokens(self, results: List[PremiseResult]):
         """
         Converts the features in the given patterns from their index representation to their token representation.

src/pypremise/core.py

@@ -35,6 +35,9 @@
 PREMISE_ENGINE_NAME_APPLE_SILICON = "AppleSilicon"
 PREMISE_LINUX_FILENAME = "Premise_Linux"
 PREMISE_APPLESILICON_FILENAME = "Premise_AppleSilicon"
+PREMISE_ENGINE_NAME_WINDOWS = "Premise_Windows"
+PREMISE_WINDOWS_FILENAME = "Premise_Windows.exe"
+
 
 logger = logging.getLogger(__name__)
 
@@ -87,35 +90,62 @@ def write_embedding_file(embedding_index_to_vector, embedding_path: str, embeddi
     out_file.flush()
 
 
-def call_premise_program(feature_path: str, label_path: str, result_path: str, embedding_path: str,
-                         embedding_size: int, neighbor_max_distance: int, fisher_p_value: float,
-                         clause_max_overlap: float, min_overlap: float, 
-                         premise_engine: Optional[str] = None) -> None:
-    premise_path = get_premise_path(premise_engine)
+def call_premise_program(
+    feature_path,
+    label_path,
+    result_path,
+    embedding_path,
+    embedding_size,
+    neighbor_max_distance,
+    fisher_p_value,
+    clause_max_overlap,
+    min_overlap,
+    premise_engine
+):
+    import subprocess, logging
+    logger = logging.getLogger(__name__)
 
     logger.info("Starting Premise. This might take a while.")
+
+    cmd = [
+        premise_engine,
+        feature_path,
+        label_path,
+        result_path,
+        embedding_path,
+        str(embedding_size),
+        str(neighbor_max_distance),
+        str(fisher_p_value),
+        str(clause_max_overlap),
+        str(min_overlap),
+    ]
+
+    logger.info("Premise command: " + " ".join(cmd))
+
     try:
-        stdout_output = ""
-        with subprocess.Popen([premise_path, feature_path, label_path, result_path,
-                                    embedding_path,
-                                    str(embedding_size), str(neighbor_max_distance),
-                                    str(fisher_p_value), str(clause_max_overlap), str(min_overlap)],
-                                   stdout=subprocess.PIPE,
-                                   stderr=subprocess.PIPE, text=True) as process:
-            for line in process.stdout:
-                logger.info(line)
-            stdout_output += line
-            process.wait()
-
-            stderr_output = process.stderr.read()
+        process = subprocess.Popen(
+            cmd,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            text=True
+        )
+
+        stdout_output, stderr_output = process.communicate()
+
+        if stdout_output:
+            logger.info("Premise stdout:\n" + stdout_output)
+        if stderr_output:
+            logger.error("Premise stderr:\n" + stderr_output)
+
+        if process.returncode != 0:
+            raise Exception(
+                f"Premise exited with code {process.returncode}. "
+                f"stderr:\n{stderr_output}"
+            )
 
     except Exception as e:
         raise Exception(f"Execution of Premise failed due to '{e}'.")
-    
-    if len(stderr_output) > 0:
-        logger.error(f"Premise reported an error.\nThe error output was" + 
-                     "\n\n-------\n{stderr}\n-------")
-        raise Exception(f"Premise reported an error.")
+
 
 
 def get_premise_path(premise_engine: Optional[str] = None):
@@ -158,6 +188,8 @@ def get_premise_path(premise_engine: Optional[str] = None):
             premise_engine = PREMISE_ENGINE_NAME_LINUX
         elif try_premise(os.path.join(module_path, PREMISE_APPLESILICON_FILENAME)):
             premise_engine = PREMISE_ENGINE_NAME_APPLE_SILICON
+        elif try_premise(os.path.join(module_path, PREMISE_WINDOWS_FILENAME)):
+            premise_engine = PREMISE_ENGINE_NAME_WINDOWS
 
         if premise_engine != None:
             # we found one! Let's try to store it
@@ -179,6 +211,8 @@ def get_premise_path(premise_engine: Optional[str] = None):
         path = os.path.join(module_path, PREMISE_LINUX_FILENAME)
     elif premise_engine == PREMISE_ENGINE_NAME_APPLE_SILICON:
         path = os.path.join(module_path, PREMISE_APPLESILICON_FILENAME)
+    elif premise_engine == PREMISE_ENGINE_NAME_WINDOWS:
+        path = os.path.join(module_path, PREMISE_WINDOWS_FILENAME)
     else:
         raise Exception(f"Unknown Premise engine '{premise_engine}'")