From 449da48076d0b9ca525dfd1b46edc5e8ae157ea8 Mon Sep 17 00:00:00 2001
From: Matthew Wells <mattwells9@shaw.ca>
Date: Thu, 4 Jul 2024 15:17:57 -0500
Subject: [PATCH] added check for required utilities

---
 locidex/constants.py |  8 ++++++++
 locidex/main.py      |  9 +++++++--
 locidex/utils.py     | 34 ++++++++++++++++++++++++++++++++--
 3 files changed, 47 insertions(+), 4 deletions(-)

diff --git a/locidex/constants.py b/locidex/constants.py
index d04074f..37b6b29 100644
--- a/locidex/constants.py
+++ b/locidex/constants.py
@@ -192,3 +192,11 @@ class LocidexDBHeader(NamedTuple):
     count_int_stops: int
     dna_ambig_count: int
 
+UTILITIES_CHECK = [
+    "makeblastdb",
+    BlastCommands.tblastn,
+    BlastCommands.blastn,
+    BlastCommands.blastp,
+    "mafft",
+    "blastdbcheck"
+]
\ No newline at end of file
diff --git a/locidex/main.py b/locidex/main.py
index 963995d..ce2f635 100644
--- a/locidex/main.py
+++ b/locidex/main.py
@@ -4,7 +4,7 @@
 import argparse
 import traceback
 import logging
-from . import format, extract, report, merge, search, build, manifest
+from . import format, extract, report, merge, search, build, manifest, utils, constants
 
 logger = logging.getLogger(__name__)
 logging.basicConfig(filemode=sys.stderr, level=logging.DEBUG)
@@ -28,14 +28,19 @@ def main(argv=None):
     for k, v in tasks.items():
         format_parser = sub_parsers.add_parser(k, description=v[help_msg], help=v[help_msg])
         v[module_idx].add_args(format_parser)
-        
+    
     args = parser.parse_args(argv)
     if args.command is None:
         parser.print_help()
+        try:
+            utils.check_utilities(logger, constants.UTILITIES_CHECK)
+        except FileNotFoundError:
+            pass
         sys.exit()
 
     error_file = "errors.txt"
     try:
+        utils.check_utilities(logger, constants.UTILITIES_CHECK)
         logger.info("Running {}".format(args.command))
         tasks[args.command][module_idx].run(args)
         logger.info("Finished: {}".format(args.command))
diff --git a/locidex/utils.py b/locidex/utils.py
index df06ee9..318dd4c 100644
--- a/locidex/utils.py
+++ b/locidex/utils.py
@@ -1,15 +1,19 @@
 import hashlib
 import json
 import os
+import shutil
 import argparse
+import logging
+import errno
 from collections import Counter
 from pathlib import Path
 from locidex.manifest import ManifestItem
 from Bio.Seq import Seq
 from typing import Dict, FrozenSet, Optional, List
-from locidex.constants import NT_SUB, PROTEIN_ALPHA, DNA_ALPHA, OPTION_GROUPS, FILE_TYPES
+from locidex.constants import NT_SUB, PROTEIN_ALPHA, DNA_ALPHA, OPTION_GROUPS, FILE_TYPES, raise_file_not_found_e
 import locidex.manifest as manifest 
 
+
 def slots(annotations: Dict[str, object]) -> FrozenSet[str]:
     """
     Thank you for this: https://stackoverflow.com/a/63658478
@@ -161,4 +165,30 @@ def get_format(file: Path) -> Optional[str]:
             if ext in extensions:
                 format = k
                 break
-    return format
\ No newline at end of file
+    return format
+
+def check_utility_installed(utility: str) -> Optional[str]:
+    """
+    utility str: The name of an executable to verify installation of
+    return None if the utility is installed | Message if utility is not installed
+    """
+    out_string: Optional[str]  = None
+    utility_path: Optional[str] = shutil.which(utility)
+    if utility_path is None:
+        out_string = f"Utility {utility} is not installed or is not executable."
+    return out_string
+
+
+def check_utilities(logger: logging.Logger, utilities: List[str]):
+    """
+    Check all utilites are installed
+    """
+    missing_utilities: List[str] = []
+    for i in utilities:
+        output = check_utility_installed(i)
+        if output is not None:
+            missing_utilities.append(i)
+            logger.critical("Missing: {}".format(output))
+    
+    if missing_utilities:
+        raise FileNotFoundError(errno.ENOENT, os.strerror(errno.ENOENT), ",".join(missing_utilities))
\ No newline at end of file