From 3ccc6720e722e2178b21ac358daa55635ecdaed2 Mon Sep 17 00:00:00 2001
From: Adam Sawicki <adam.sawicki@ibm.com>
Date: Wed, 8 Jan 2020 01:13:59 -0500
Subject: [PATCH 1/2] Add '--no-logs' option to argument parser

---
 rnlp/__main__.py | 20 +++++++++++---------
 1 file changed, 11 insertions(+), 9 deletions(-)

diff --git a/rnlp/__main__.py b/rnlp/__main__.py
index 085c63e..208beaa 100644
--- a/rnlp/__main__.py
+++ b/rnlp/__main__.py
@@ -38,15 +38,6 @@
 LOGGER = logging.getLogger(__name__)
 LOGGER.setLevel(logging.INFO)
 
-LOG_HANDLER = logging.FileHandler("rnlp_log.log")
-LOG_HANDLER.setLevel(logging.INFO)
-FORMATTER = logging.Formatter("%(asctime)s - %(name)s - %(levelname)s - %(message)s")
-LOG_HANDLER.setFormatter(FORMATTER)
-
-LOGGER.addHandler(LOG_HANDLER)
-LOGGER.info("Started logger.")
-
-LOGGER.info("Started Argument Parser.")
 PARSER = argparse.ArgumentParser(
     description="rnlp (v{0}): Convert text into relational facts.".format(__version__),
     epilog="This program is free software under the {0}. {1}".format(
@@ -61,8 +52,19 @@
     "-d", "--directory", type=str, help="Read all .txt files in directory"
 )
 FILE_OR_DIR.add_argument("-f", "--file", type=str, help="Read from one .txt file")
+PARSER.add_argument("--no-logs", action="store_true", help="Specify that no logs should be created.")
+
 
 ARGS = PARSER.parse_args()
+LOG_HANDLER = logging.NullHandler() if ARGS.no_logs else logging.FileHandler("rnlp_log.log")
+
+LOG_HANDLER.setLevel(logging.INFO)
+FORMATTER = logging.Formatter("%(asctime)s - %(name)s - %(levelname)s - %(message)s")
+LOG_HANDLER.setFormatter(FORMATTER)
+
+LOGGER.addHandler(LOG_HANDLER)
+LOGGER.info("Started logger.")
+
 LOGGER.info("Argument Parsing Successful.")
 
 N_BLOCKS = ARGS.blockSize

From f3292d529f692b059918a1d0042888d06e95e7e7 Mon Sep 17 00:00:00 2001
From: Adam Sawicki <adam.sawicki@ibm.com>
Date: Fri, 24 Jan 2020 12:23:47 -0500
Subject: [PATCH 2/2] Add check_state module to ensure proper nltk setup

---
 rnlp/check_state.py    | 54 ++++++++++++++++++++++++++++++++++++++++++
 rnlp/parse.py          |  2 ++
 rnlp/textprocessing.py |  3 +++
 3 files changed, 59 insertions(+)
 create mode 100644 rnlp/check_state.py

diff --git a/rnlp/check_state.py b/rnlp/check_state.py
new file mode 100644
index 0000000..9414723
--- /dev/null
+++ b/rnlp/check_state.py
@@ -0,0 +1,54 @@
+_err = (
+    """
+    >>> import nltk
+    >>> nltk.download('punkt')
+    >>> nltk.download('stopwords')
+    >>> nltk.download('averaged_perceptron_tagger')
+
+    Visit https://rnlp.readthedocs.io/en/latest/getting_started/02_installation.html
+    for more information.
+    """
+)
+
+
+def _ensure_nltk_installed():
+    """
+    Determine if `nltk` is installed.
+
+    :raises: Exception if `nltk` is not installed.
+    """
+    try:
+        import nltk
+    except ModuleNotFoundError:
+        raise Exception(
+            "Unable to `import nltk` because it is not in the current environment."
+            " Run `pip install nltk`, and then the following in an interpreter:\n"
+            + _err
+        ) from  None
+
+
+def _find_nltk_module(module_name):
+    """
+    Determine whether a certain `nltk` module is installed.
+
+    :param module_name: Name of a module within nltk
+    :type module name: str.
+
+    :raises: Exception if the module cannot be found
+    """
+    import nltk
+    try:
+        nltk.data.find(module_name)
+    except LookupError as e:
+        raise Exception(
+            "Unable to find module '" + module_name + "'. Please run the following:\n"
+            + _err
+        ) from None
+
+
+def ensure_nltk_setup():
+    """Function to ensure required packages are installed."""
+    _ensure_nltk_installed()
+    nltk_modules = ("tokenizers/punkt", "corpora/stopwords", "taggers/averaged_perceptron_tagger")
+    for nltk_module in nltk_modules:
+        _find_nltk_module(nltk_module)
diff --git a/rnlp/parse.py b/rnlp/parse.py
index 66ca937..8bee26a 100644
--- a/rnlp/parse.py
+++ b/rnlp/parse.py
@@ -22,6 +22,8 @@
 rnlp.parse
 ----------
 """
+import rnlp.check_state as check_state
+check_state.ensure_nltk_setup()
 
 import string
 import nltk
diff --git a/rnlp/textprocessing.py b/rnlp/textprocessing.py
index 5fa741a..d68ed89 100644
--- a/rnlp/textprocessing.py
+++ b/rnlp/textprocessing.py
@@ -37,6 +37,9 @@
 The depth of reasoning probably depends on the domain you are working on.
 """
 
+import rnlp.check_state as check_state
+check_state.ensure_nltk_setup()
+
 import string
 from nltk import sent_tokenize
 from nltk.corpus import stopwords