Merge pull request jxmorris12#86 from pidefrem/deactivate_download_if_env_var_defined

Introduce new env var `LTP_JAR_DIR_PATH` to let the user specify a custom installation path of LT

Author: jxmorris12

README.md

@@ -4,7 +4,7 @@
 
 ![Test with PyTest](https://github.com/jxmorris12/language_tool_python/workflows/Test%20with%20PyTest/badge.svg)
 
-Current LanguageTool version: **5.5**
+Current LanguageTool version: **6.4**
 
 This is a Python wrapper for [LanguageTool](https://languagetool.org). LanguageTool is open-source grammar tool, also known as the spellchecker for OpenOffice. This library allows you to make to detect grammar errors and spelling mistakes through a Python script or through a command-line interface.
 
@@ -151,7 +151,7 @@ You can run LanguageTool on one host and connect to it from another.  This is us
 
 ## Configuration
 
-LanguageTool offers lots of built-in configuration options. 
+LanguageTool offers lots of built-in configuration options.
 
 ### Example: Enabling caching
 Here's an example of using the configuration options to enable caching. Some users have reported that this helps performance a lot.
@@ -222,13 +222,20 @@ Searching for a specific rule to enable or disable? Curious the breadth of rules
 
 ### Customizing Download URL or Path
 
+If LanguageTool is already installed on your system, you can defined the following environment variable:
+```bash
+$ export LTP_JAR_DIR_PATH = /path/to/the/language/tool/jar/files
+```
+
+Overwise, `language_tool_python` can download LanguageTool for you automatically.
+
 To overwrite the host part of URL that is used to download LanguageTool-{version}.zip:
 
 ```bash
 $ export LTP_DOWNLOAD_HOST = [alternate URL]
 ```
 
-This can be used to downgrade to an older version, for example, or to download from a mirror. 
+This can be used to downgrade to an older version, for example, or to download from a mirror.
 
 And to choose the specific folder to download the server to:
 
@@ -252,6 +259,7 @@ into where the ``language_tool_python`` package resides.
 
 As of April 2020, `language_tool_python` was forked from `language-check` and no longer supports LanguageTool versions lower than 4.0.
 
-### Acknowledgements 
+### Acknowledgements
+
 This is a fork of https://github.com/myint/language-check/ that produces more easily parsable
 results from the command-line.

language_tool_python/__main__.py

@@ -8,7 +8,7 @@
 from .server import LanguageTool
 from .utils import LanguageToolError
 
-import pkg_resources 
+import pkg_resources
 __version__ = pkg_resources.require("language_tool_python")[0].version
 
 
@@ -175,4 +175,4 @@ def main():
     return status
 
 
-sys.exit(main())
+sys.exit(main())

language_tool_python/download_lt.py

@@ -2,7 +2,6 @@
 # -*- coding: utf-8 -*-
 """Download latest LanguageTool distribution."""
 
-import glob
 import logging
 import os
 import re
@@ -11,11 +10,16 @@
 import sys
 import tempfile
 import tqdm
+from typing import Optional
 import zipfile
 
 from distutils.spawn import find_executable
 from urllib.parse import urljoin
-from .utils import get_language_tool_download_path
+from .utils import (
+    find_existing_language_tool_downloads,
+    get_language_tool_download_path,
+    LTP_JAR_DIR_PATH_ENV_VAR
+)
 
 # Create logger for this file.
 logging.basicConfig(format='%(message)s')
@@ -27,7 +31,7 @@
 BASE_URL = os.environ.get('LTP_DOWNLOAD_HOST', 'https://www.languagetool.org/download/')
 FILENAME = 'LanguageTool-{version}.zip'
 
-LATEST_VERSION = '6.2'
+LTP_DOWNLOAD_VERSION = '6.4'
 
 JAVA_VERSION_REGEX = re.compile(
     r'^(?:java|openjdk) version "(?P<major1>\d+)(|\.(?P<major2>\d+)\.[^"]+)"',
@@ -56,27 +60,36 @@ def parse_java_version(version_text):
     (1, 8)
 
     """
-    match = re.search(JAVA_VERSION_REGEX, version_text) or re.search(JAVA_VERSION_REGEX_UPDATED, version_text)
+    match = (
+        re.search(JAVA_VERSION_REGEX, version_text)
+        or re.search(JAVA_VERSION_REGEX_UPDATED, version_text)
+    )
     if not match:
         raise SystemExit(
             'Could not parse Java version from """{}""".'.format(version_text))
     major1 = int(match.group('major1'))
     major2 = int(match.group('major2')) if match.group('major2') else 0
     return (major1, major2)
 
+
 def confirm_java_compatibility():
     """ Confirms Java major version >= 8. """
     java_path = find_executable('java')
     if not java_path:
-        raise ModuleNotFoundError('No java install detected. Please install java to use language-tool-python.')
+        raise ModuleNotFoundError(
+            'No java install detected. '
+            'Please install java to use language-tool-python.'
+        )
 
     output = subprocess.check_output([java_path, '-version'],
                                      stderr=subprocess.STDOUT,
                                      universal_newlines=True)
 
     major_version, minor_version = parse_java_version(output)
-    # Some installs of java show the version number like `14.0.1` and others show `1.14.0.1`
-    # (with a leading 1). We want to support both, as long as the major version is >= 8.
+    # Some installs of java show the version number like `14.0.1`
+    # and others show `1.14.0.1`
+    # (with a leading 1). We want to support both,
+    # as long as the major version is >= 8.
     # (See softwareengineering.stackexchange.com/questions/175075/why-is-java-version-1-x-referred-to-as-java-x)
     if major_version == 1 and minor_version >= 8:
         return True
@@ -85,31 +98,37 @@ def confirm_java_compatibility():
     else:
         raise SystemError('Detected java {}.{}. LanguageTool requires Java >= 8.'.format(major_version, minor_version))
 
+
 def get_common_prefix(z):
     """Get common directory in a zip file if any."""
     name_list = z.namelist()
     if name_list and all(n.startswith(name_list[0]) for n in name_list[1:]):
         return name_list[0]
     return None
 
+
 def http_get(url, out_file, proxies=None):
     """ Get contents of a URL and save to a file.
     """
     req = requests.get(url, stream=True, proxies=proxies)
     content_length = req.headers.get('Content-Length')
     total = int(content_length) if content_length is not None else None
-    if req.status_code == 403: # Not found on AWS
+    if req.status_code == 403:  # Not found on AWS
         raise Exception('Could not find at URL {}.'.format(url))
-    progress = tqdm.tqdm(unit="B", unit_scale=True, total=total, desc=f'Downloading LanguageTool {LATEST_VERSION}')
+    progress = tqdm.tqdm(unit="B", unit_scale=True, total=total,
+                         desc=f'Downloading LanguageTool {LTP_DOWNLOAD_VERSION}')
     for chunk in req.iter_content(chunk_size=1024):
-        if chunk: # filter out keep-alive new chunks
+        if chunk:  # filter out keep-alive new chunks
             progress.update(len(chunk))
             out_file.write(chunk)
     progress.close()
 
+
 def unzip_file(temp_file, directory_to_extract_to):
     """ Unzips a .zip file to folder path. """
-    logger.info('Unzipping {} to {}.'.format(temp_file.name, directory_to_extract_to))
+    logger.info(
+        'Unzipping {} to {}.'.format(temp_file.name, directory_to_extract_to)
+    )
     with zipfile.ZipFile(temp_file.name, 'r') as zip_ref:
         zip_ref.extractall(directory_to_extract_to)
 
@@ -128,26 +147,34 @@ def download_zip(url, directory):
     # Tell the user the download path.
     logger.info('Downloaded {} to {}.'.format(url, directory))
 
-def download_lt():
-    download_folder = get_language_tool_download_path()
-    assert os.path.isdir(download_folder)
-    old_path_list = [
-        path for path in
-        glob.glob(os.path.join(download_folder, 'LanguageTool*'))
-        if os.path.isdir(path)
-    ]
 
+def download_lt(language_tool_version: Optional[str] = LTP_DOWNLOAD_VERSION):
     confirm_java_compatibility()
-    version = LATEST_VERSION
-    filename = FILENAME.format(version=version)
-    language_tool_download_url = urljoin(BASE_URL, filename)
-    dirname, _ = os.path.splitext(filename)
-    extract_path = os.path.join(download_folder, dirname)
 
-    if extract_path in old_path_list:
+    download_folder = get_language_tool_download_path()
+
+    # Use the env var to the jar directory if it is defined
+    # otherwise look in the download directory
+    if os.environ.get(LTP_JAR_DIR_PATH_ENV_VAR):
         return
 
-    download_zip(language_tool_download_url, download_folder)
+    # Make download path, if it doesn't exist.
+    os.makedirs(download_folder, exist_ok=True)
+
+    assert os.path.isdir(download_folder)
+    old_path_list = find_existing_language_tool_downloads(download_folder)
+
+    if language_tool_version:
+        version = language_tool_version
+        filename = FILENAME.format(version=version)
+        language_tool_download_url = urljoin(BASE_URL, filename)
+        dirname, _ = os.path.splitext(filename)
+        extract_path = os.path.join(download_folder, dirname)
+
+        if extract_path in old_path_list:
+            return
+        download_zip(language_tool_download_url, download_folder)
+
 
 if __name__ == '__main__':
     sys.exit(download_lt())