feat: health check, install binaries, font etc.

Author: kiyoon

.github/workflows/tests.yml

@@ -42,6 +42,7 @@ jobs:
           # python -m pip install --upgrade pip
           uv pip install -r deps/requirements_dev.txt
           uv pip install -e .
+          bash scripts/install_binaries.sh
       - name: Run pytest
         run: |
           set +e  # Do not exit shell on pytest failure
@@ -107,6 +108,7 @@ jobs:
           # python -m pip install --upgrade pip
           uv pip install -r deps/requirements_dev.txt
           uv pip install -e .
+          bash scripts/install_binaries.sh
       - name: Run doctest
         run: |
           set +e  # Do not exit shell on pytest failure

README.md

@@ -22,13 +22,14 @@
 ## 돌려 보기
 
 1. (Optional) `pip3 install --user uv` 해서 pip 대신 `uv pip` 사용하면 더 빠름.
-2. `uv pip install -r deps/requirements.txt`, `uv pip install -e .`으로 dependencies 및 mlproject 패키지 설치
+2. `uv pip install -r deps/requirements.txt`, `uv pip install -e .`, `bash scripts/install_binaries.sh` 으로 dependencies 및 mlproject 패키지 설치
 3. template.env 파일을 .env로 복사한 후 token 등 내용 수정.
-4. `python tools/examples/color_logging_main.py` 실행해보기. 로깅 내용은 `data/logs` 폴더 안에 기록됨.
-5. `uv pip install -r deps/requirements_dev.txt` 으로 pytest 등 개발자용 패키지도 설치가능
-6. `pytest` 커맨드로 테스트 실행해보기.
+4. `python -m mlproject.health` 실행해서 환경 설정이 잘 되었는지 확인.
+5. `python tools/examples/color_logging_main.py` 실행해보기. 로깅 내용은 `data/logs` 폴더 안에 기록됨.
+6. `uv pip install -r deps/requirements_dev.txt` 으로 pytest 등 개발자용 패키지도 설치가능
+7. `pytest` 커맨드로 테스트 실행해보기.
     - `python scripts/run_doctest.py` 커맨드로 doctest 실행해보기.
-7. `import mlproject; print(mlproject.__version__)` 해보면 `0.1.0+4.g75bbed7.dirty` 이런식으로 나옴.  
+8. `import mlproject; print(mlproject.__version__)` 해보면 `0.1.0+4.g75bbed7.dirty` 이런식으로 나옴.  
     - 0.1.0 버전 이후 4개의 커밋이란 뜻. 그리고 커밋되지 않은 수정사항이 있는 상태이면 dirty버전임.
 
 ## 파일 설명

scripts/health.py

@@ -0,0 +1,48 @@
+#!/usr/bin/env bash
+
+if [[ -n "$CONDA_PREFIX" ]]; then
+	PREFIX=$CONDA_PREFIX
+else
+	echo "Please activate a conda environment"
+	exit 1
+fi
+
+if command -v mamba &> /dev/null; then
+	CONDA=mamba
+else
+	CONDA=conda
+fi
+
+echo "Installing binaries to $PREFIX"
+TEMPDIR=$(mktemp -d)
+
+# Shared objects needed for the cairosvg package
+$CONDA install -y -c conda-forge cairo
+
+# Install FiraCode font (for rich export_svg -> cairosvg PDF generation)
+if [[ "$OSTYPE" == "darwin"* ]]; then
+	brew tap homebrew/cask-fonts
+	brew install font-fira-code
+else
+	FONTDIR="$HOME/.local/share/fonts"
+	mkdir -p "$FONTDIR"
+
+	if [ ! command -v fc-list ] &>/dev/null || ! fc-list | grep -q "FiraCode"; then
+		echo "FiraCode could not be found. Installing on $FONTDIR"
+		# NOTE: we need non-NF FiraCode for python rich export_svg -> cairosvg PDF generation
+		mkdir -p "$TEMPDIR/firacode"
+		curl -s https://api.github.com/repos/tonsky/FiraCode/releases/latest |
+			grep "browser_download_url.*.zip" |
+			cut -d : -f 2,3 |
+			tr -d \" |
+			wget -qi - -O "$TEMPDIR/firacode.zip"
+		unzip "$TEMPDIR/firacode.zip" -d "$TEMPDIR"/firacode
+		mv "$TEMPDIR"/firacode/ttf/*.ttf "$FONTDIR"
+
+		fc-cache -fv
+	else
+		echo "FiraCode is already installed"
+	fi
+fi
+
+rm -rf "$TEMPDIR"

scripts/install_binaries.sh

@@ -1,8 +1,12 @@
 import os
 from pathlib import Path
 
+from dotenv import load_dotenv
+
 from . import _version
 
+load_dotenv()
+
 __version__ = _version.get_versions()["version"]
 
 default_log_level = os.environ.get("MLPROJECT_LOG_LEVEL")

src/mlproject/__init__.py

@@ -1,15 +1,14 @@
 """Check health of the installation."""
 
+import logging
 import os
 
-from rich import print
-from rich.console import Console
-
 from .. import DATA_DIR
+from .font import verify_fonts_bool
 from .slack import check_env as slack_check_env
 from .slack import check_send_text as slack_check_send_text
 
-console = Console()
+logger = logging.getLogger(__name__)
 
 
 def check_binaries():
@@ -19,16 +18,16 @@ def check_binaries():
 
 def check_env():
     """Check environment variables."""
-    data_dir = os.environ.get("MLPROJECT_DATA_DIR")
+    ppmi_data_dir = os.environ.get("MLPROJECT_DATA_DIR")
 
-    if data_dir is None or data_dir == "":
-        print(
-            "🤢 Please set the environment variable MLPROJECT_DATA_DIR to the path of the data directory."
+    if ppmi_data_dir is None:
+        logger.warning(
+            "🤒 Please set the environment variable MLPROJECT_DATA_DIR to the path of the data directory.\n"
+            f"Otherwise, the default {DATA_DIR} will be used."
         )
-        print(f"Otherwise, the default {DATA_DIR} will be used.")
         return False
 
-    print(f"✅ MLPROJECT_DATA_DIR is set to {data_dir}")
+    logger.info(f"✅ MLPROJECT_DATA_DIR is set to {ppmi_data_dir}")
     return True
 
 
@@ -39,9 +38,11 @@ def main():
     successes.append(slack_check_env())
     successes.append(slack_check_send_text())
 
+    successes.append(verify_fonts_bool())
+
     if all(successes):
-        print()
-        print("💪 You are ready to go!")
+        logger.info("")
+        logger.info("💪 You are ready to go!")
 
 
 if __name__ == "__main__":

src/mlproject/health/__init__.py

@@ -0,0 +1,6 @@
+from ..utils import setup_logging
+from . import main
+
+if __name__ == "__main__":
+    setup_logging(output_files=[], file_levels=[])
+    main()

src/mlproject/health/__main__.py

@@ -0,0 +1,43 @@
+import logging
+import subprocess
+
+logger = logging.getLogger(__name__)
+
+
+class FontNotInstalledError(Exception):
+    pass
+
+
+def verify_fonts_installed():
+    """
+    This function verifies that the FiraCode font is installed because Rich Console.export_svg() uses it.
+
+    Actually, the svg itself doesn't need the font but when we convert it to PDF using cairosvg, the font is needed.
+    """
+    cmd = ["fc-match", "FiraCode:style=Regular"]
+    output = subprocess.Popen(cmd, stdout=subprocess.PIPE).communicate()[0]
+
+    if not output.decode("utf-8").startswith("FiraCode-Regular"):
+        raise FontNotInstalledError("FiraCode")
+
+    cmd = ["fc-match", "FiraCode:style=Bold"]
+    output = subprocess.Popen(cmd, stdout=subprocess.PIPE).communicate()[0]
+
+    if not output.decode("utf-8").startswith("FiraCode-Bold"):
+        raise FontNotInstalledError("FiraCode")
+
+    # log.debug('Courier and Helvetica Neue were found to be installed.')
+
+
+def verify_fonts_bool():
+    try:
+        verify_fonts_installed()
+        logger.info("✅ FiraCode font is installed.")
+        return True
+    except FontNotInstalledError:
+        logger.error("😡 FiraCode font is not installed.")
+        return False
+
+
+if __name__ == "__main__":
+    verify_fonts_installed()

src/mlproject/health/font.py

@@ -1,41 +1,55 @@
-# ruff: noqa: T201
+import logging
 import os
 
 from ..utils.slack.send_only import send_text
 
+logger = logging.getLogger(__name__)
 
-def check_one_env(env_var):
+
+def check_one_env(env_var, secret=False, secret_show_first_n=0):
     value = os.environ.get(env_var)
     if value is None:
-        print(f"🤢 Please set the environment variable {env_var}.")
+        logger.error(f"😡 Please set the environment variable {env_var}.")
         return False
-    print(f"✅ {env_var} is set to {value}")
+
+    if secret:
+        if secret_show_first_n > 0:
+            logger.info(
+                f"✅ {env_var} is set to {value[:secret_show_first_n]}{'*' * (len(value) - secret_show_first_n)}"
+            )
+        else:
+            logger.info(f"✅ {env_var} is set.")
+    else:
+        logger.info(f"✅ {env_var} is set to {value}")
     return True
 
 
-def check_many_env(env_vars):
+def check_many_env(env_vars, secret=False, secret_show_first_n=0):
     ret = True
     for env_var in env_vars:
-        if not check_one_env(env_var):
+        if not check_one_env(env_var, secret, secret_show_first_n):
             ret = False
     return ret
 
 
 def check_env():
-    return check_many_env(["SLACK_BOT_TOKEN", "SLACK_APP_TOKEN", "SLACK_CHANNEL_ID"])
+    secrets_checked = check_many_env(["SLACK_BOT_TOKEN", "SLACK_APP_TOKEN"], True, 5)
+    normal_checked = check_many_env(["SLACK_CHANNEL_ID"])
+
+    return secrets_checked and normal_checked
 
 
 def check_send_text():
-    print("🚀 Sending a test message to the Slack channel...")
+    logger.info("🚀 Sending a test message to the Slack channel...")
     response = send_text("💪 Checking health of the Slack bot.")
 
     if response is None:
-        print("🤢 The response is None.")
+        logger.error("😡 The response is None.")
         return False
 
     if response["ok"]:
-        print("✅ The message was sent successfully.")
+        logger.info("✅ The message was sent successfully.")
         return True
 
-    print(f"🤢 The message was not sent successfully. Response: {response}")
+    logger.error(f"😡 The message was not sent successfully. Response: {response}")
     return False

src/mlproject/health/slack.py

@@ -233,12 +233,24 @@ def send_svg_as_pdf(
         channel_id = default_channel_id
         assert channel_id is not None
 
+    # svg_buf = BytesIO()
     pdf_buf = BytesIO()
 
     if isinstance(svg_file, str | bytes):
         cairosvg.svg2pdf(bytestring=svg_file, write_to=pdf_buf)
+        # raise NotImplementedError("cairosvg does not support bytestring")
+        # str to bytes
+        # if isinstance(svg_file, str):
+        #     svg_file = svg_file.encode("utf-8")
+        # svg_buf.write(svg_file)
+        # svg_buf.seek(0)
+        # drawing = svg2rlg(svg_buf)
+        # renderPDF.drawToFile(drawing, pdf_buf)
+
     elif isinstance(svg_file, IOBase):
-        cairosvg.svg2pdf(file_obj=svg_file, write_to=pdf_buf)
+        # cairosvg.svg2pdf(file_obj=svg_file, write_to=pdf_buf)
+        drawing = svg2rlg(svg_file)
+        renderPDF.drawToFile(drawing, pdf_buf)
     else:
         raise ValueError(f"Unsupported type {type(svg_file)}")