Merge pull request #5 from andrewalkermo/main

versão inicial documentada do DPToie-Python

Author: andrewalkermo

.dockerignore

@@ -0,0 +1,25 @@
+# Arquivo: .dockerignore
+
+# Ignorar pastas de ambiente virtual
+.venv/
+venv/
+env/
+
+# Ignorar cache do Python
+__pycache__/
+*.pyc
+
+# Ignorar cache e pastas do Git
+.git/
+.gitignore
+
+# Ignorar arquivos do Docker
+Dockerfile
+.dockerignore
+
+# Outros
+.idea/
+.vscode/
+dist/
+build/
+*.egg-info/

.editorconfig

@@ -0,0 +1,16 @@
+root = true
+
+[*]
+charset = utf-8
+end_of_line = lf
+indent_size = 4
+indent_style = space
+insert_final_newline = false
+max_line_length = 120
+tab_width = 4
+
+[*.yml]
+indent_size = 2
+
+[*.json]
+indent_size = 2

.github/workflows/extrai_fatos.yml

@@ -0,0 +1,84 @@
+name: extrai_fatos
+
+on: [push, workflow_dispatch]
+
+jobs:
+  extrai_fatos:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Build container
+        run: DOCKER_BUILDKIT=1 docker build --tag ptoie_python .
+
+      - name: Run all inputs with all configurations
+        run: |
+          set -euo pipefail
+          mkdir -p outputs
+
+          # Define flag combinations to run (same for JSON and CSV)
+          FLAGS_LIST=("" "-cc" "-sc" "-hs" "-a" "-a -t" "-cc -sc -hs -a -t")
+
+          # iterate only over .conll files in the inputs directory
+          for input_path in inputs/*.conll; do
+            [ -f "$input_path" ] || continue
+            filename=$(basename "$input_path")
+            # strip last extension only (handles names like ceten-200.conll)
+            name="${filename%.*}"
+
+            # detect input type: .txt -> --input-type txt, otherwise assume conll
+            if [[ "$filename" == *.txt ]]; then
+              input_args="--input-type txt"
+            else
+              input_args="-it conll"
+            fi
+
+            for flags in "${FLAGS_LIST[@]}"; do
+              # --- JSON ---
+              echo "[JSON] input=$filename flags=[$flags]"
+              # remove stale json output if any
+              rm -f ./outputs/output.json
+              docker run --rm -v "$(pwd)":/ptoie_python ptoie_python poetry run python3 src/main.py -i "$input_path" $input_args $flags || echo "json-run-failed for $filename $flags"
+
+              # compute cleaned suffix for filename (empty => base name only)
+              if [[ -z "$flags" ]]; then
+                out_json="./outputs/${name}.json"
+              else
+                cleaned=$(echo "$flags" | sed -E "s/^\s*-//; s/\s+-/ /g; s/\s+/-/g; s/^-//")
+                # remove any leading/trailing hyphens/spaces
+                cleaned=$(echo "$cleaned" | sed -E 's/^[ -]+//; s/[ -]+$//; s/\s+/-/g')
+                out_json="./outputs/${name}-${cleaned}.json"
+              fi
+              if [[ -f ./outputs/output.json ]]; then
+                mv ./outputs/output.json "$out_json"
+              else
+                echo "Warning: JSON output not produced for $filename with flags [$flags]"
+              fi
+
+              # --- CSV ---
+              echo "[CSV] input=$filename flags=[$flags]"
+              # remove stale csv output if any
+              rm -f ./outputs/output.csv
+              docker run --rm -v "$(pwd)":/ptoie_python ptoie_python poetry run python3 src/main.py -i "$input_path" $input_args $flags -o ./outputs/output.csv -ot csv || echo "csv-run-failed for $filename $flags"
+
+              if [[ -z "$flags" ]]; then
+                out_csv="./outputs/${name}.csv"
+              else
+                cleaned=$(echo "$flags" | sed -E "s/^\s*-//; s/\s+-/ /g; s/\s+/-/g; s/^-//")
+                cleaned=$(echo "$cleaned" | sed -E 's/^[ -]+//; s/[ -]+$//; s/\s+/-/g')
+                out_csv="./outputs/${name}-${cleaned}.csv"
+              fi
+              if [[ -f ./outputs/output.csv ]]; then
+                mv ./outputs/output.csv "$out_csv"
+              else
+                echo "Warning: CSV output not produced for $filename with flags [$flags]"
+              fi
+
+            done
+          done
+
+      - name: Upload artifacts
+        uses: actions/upload-artifact@v4
+        with:
+          name: extrai_fatos_outputs
+          path: |
+            ./outputs

.gitignore

@@ -0,0 +1,8 @@
+.idea
+.vscode
+.hypothesis
+out.txt
+/outputs/*
+!/outputs/.gitkeep
+/models/*
+!/models/.gitkeep

Dockerfile

@@ -0,0 +1,14 @@
+FROM python:3.12-slim
+
+WORKDIR /ptoie_python
+
+COPY pyproject.toml poetry.lock /ptoie_python/
+RUN pip install poetry  \
+    && poetry config virtualenvs.create false \
+    && poetry install --only main --no-root --no-directory
+COPY . /ptoie_python
+RUN poetry install --only main
+
+ENV PYTHONPATH="$PYTHONPATH:/ptoie_python"
+
+CMD [ "poetry", "run", "python3", "src/noie.py" ]