fix: TypeError: Object with dtype category cannot perform the NumPy op log1p (#108)

* fix: TypeError - Object with dtype category cannot perform the numpy op log1p

Signed-off-by: HimanshuRRai <himanshu.rai@fujitsu.com>

* fix: resolve isort and pysen version compatibility issue in CI workflow

Signed-off-by: HimanshuRRai <himanshu.rai@fujitsu.com>

* fix: Upgrade upload-artifact and download-artifact to v4

Signed-off-by: HimanshuRRai <himanshu.rai@fujitsu.com>

* fix: Conflict issue Unable to upload to same artifact name

Signed-off-by: HimanshuRRai <himanshu.rai@fujitsu.com>

---------

Signed-off-by: HimanshuRRai <himanshu.rai@fujitsu.com>
Co-authored-by: HimanshuRRai <himanshu.rai@fujitsu.com>

Author: Shakeer1024-RBA

.github/workflows/lint.yml

@@ -42,7 +42,7 @@ jobs:
           python-version: ${{ env.PYTHON_VERSION }}
 
       - name: Install Dependencies
-        run: pip install pysen flake8 black isort
+        run: pip install pysen flake8 black isort==5.12.0
 
       - name: Pysen run lint
         run: pysen run lint

.github/workflows/release.yml

@@ -58,12 +58,14 @@ jobs:
         run: poetry run -- pytest -k ${{ matrix.test }}
 
       - name: Upload Coverage
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name:  ${{ matrix.test }}
           include-hidden-files: true
           path: .coverage
           retention-days: 1
+          overwrite: true
+          
 
   additional_test:
     strategy:
@@ -101,12 +103,13 @@ jobs:
         run: poetry run -- pytest -k ${{ matrix.test }}
 
       - name: Upload Coverage
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name:  ${{ matrix.test }}
           include-hidden-files: true
           path: .coverage
           retention-days: 1
+          overwrite: true
 
   report_coverage:
     runs-on: ubuntu-latest
@@ -123,7 +126,7 @@ jobs:
           python-version: 3.11
 
       - name: Download Coverage Files
-        uses: actions/download-artifact@v3
+        uses: actions/download-artifact@v4
 
       - name: Install coverage
         run: pip install coverage

.github/workflows/test.yml

@@ -57,12 +57,13 @@ jobs:
         run: poetry run -- pytest -k ${{ matrix.test }}
 
       - name: Upload Coverage
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name:  ${{ matrix.test }}
           include-hidden-files: true
           path: .coverage
           retention-days: 1
+          overwrite: true
 
   report_coverage:
     runs-on: ubuntu-latest
@@ -78,7 +79,7 @@ jobs:
           python-version: 3.11
 
       - name: Download Coverage Files
-        uses: actions/download-artifact@v3
+        uses: actions/download-artifact@v4
 
       - name: Install coverage
         run: pip install coverage

pyproject.toml

@@ -65,7 +65,7 @@ default = "sapientml_core.preprocess.default:DefaultPreprocess"
 sample-dataset = "sapientml_core.datastore.localfile.export_modules"
 
 [tool.pysen]
-version = "0.10"
+version = "0.11.0"
 
 [tool.pysen-cli]
 settings_dir = ".pysen"

sapientml_core/adaptation/generation/pipeline_template.py

@@ -17,6 +17,7 @@
 import sys
 from pathlib import Path
 
+import pandas as pd
 from jinja2 import Environment, FileSystemLoader
 from pydantic import BaseModel
 from sapientml import macros
@@ -321,6 +322,12 @@ def generate_snippet(self, component, training_dataframe, test_dataframe, train_
             relevant_cols = sorted(list(set(relevant_cols) - set(pipeline.task.target_columns)))
         api_label = component.label_name.split(":")[2]
 
+        # Ensure only numerical columns are selected before parsing to 'log' transformation.
+        if api_label == "log":
+            relevant_cols = [
+                col for col in relevant_cols if pd.api.types.is_numeric_dtype(pipeline.all_columns_datatypes[col])
+            ]
+
         if len(relevant_cols) == 0 and api_label not in ["STANDARD", "SMOTE"]:
             logger.debug("No relevant columns found for " + api_label)
             return adapted_snippets