From ab71b34b97f3f1eee52b5688f41644541535f281 Mon Sep 17 00:00:00 2001
From: Frederick Muriuki Muriithi
Date: Mon, 5 Feb 2024 07:00:01 +0300
Subject: Do general bundle validation and show errors

* Display any and all errors on the UI
* Move `validate_bundle` to QC module and refactor to use
  `missing_files`
---
 r_qtl/errors.py    |  6 ++++++
 r_qtl/r_qtl2.py    | 22 ----------------------
 r_qtl/r_qtl2_qc.py |  9 +++++++++
 3 files changed, 15 insertions(+), 22 deletions(-)

(limited to 'r_qtl')

diff --git a/r_qtl/errors.py b/r_qtl/errors.py
index 20c5ced..417eb58 100644
--- a/r_qtl/errors.py
+++ b/r_qtl/errors.py
@@ -5,3 +5,9 @@ class RQTLError(Exception):
 
 class InvalidFormat(RQTLError):
     """Raised when the format of the file(s) is invalid."""
+
+class MissingFileError(InvalidFormat):
+    """
+    Raise when at least one file listed in the control file is missing from the
+    R/qtl2 bundle.
+    """
diff --git a/r_qtl/r_qtl2.py b/r_qtl/r_qtl2.py
index c2828d2..f8c08d9 100644
--- a/r_qtl/r_qtl2.py
+++ b/r_qtl/r_qtl2.py
@@ -288,28 +288,6 @@ def sex_information(zfile: ZipFile, cdata: dict) -> Iterator[dict]:
             key: chain(value, partial(replace_sex_info, cdata=cdata))
             for key, value in row.items() if key not in ci_fields}
 
-def validate_bundle(zfile: ZipFile):
-    """Ensure the R/qtl2 bundle is valid."""
-    cdata = control_data(zfile)
-    def __member_exists_p__(zfile, member):
-        if isinstance(member, str):
-            zfile.getinfo(member)
-        else:
-            for inner in member:
-                zfile.getinfo(inner)
-
-    try:
-        for member in (key for key in cdata.keys() if key in __FILE_TYPES__):
-            __member_exists_p__(zfile, cdata[member])
-
-        if "file" in cdata.get("sex", {}):
-            __member_exists_p__(zfile, cdata["sex"]["file"])
-
-        if "file" in cdata.get("cross_info", {}):
-            __member_exists_p__(zfile, cdata["cross_info"]["file"])
-    except KeyError as kerr:
-        raise InvalidFormat(*kerr.args) from kerr
-
 def genotype_data(zfile: ZipFile):
     """Convenience function to genotype data from R/qtl2 bundle."""
     cdata = control_data(zfile)
diff --git a/r_qtl/r_qtl2_qc.py b/r_qtl/r_qtl2_qc.py
index f666f40..261d300 100644
--- a/r_qtl/r_qtl2_qc.py
+++ b/r_qtl/r_qtl2_qc.py
@@ -3,6 +3,7 @@ from zipfile import ZipFile
 from functools import reduce
 from typing import Union, Sequence
 
+from r_qtl import errors as rqe
 from r_qtl import r_qtl2 as rqtl2
 from r_qtl.r_qtl2 import __FILE_TYPES__
 
@@ -45,3 +46,11 @@ def missing_files(zfile: ZipFile) -> tuple[str]:
 
     return tuple(filter(__missing_p__,
                         bundle_files_list(zfile, rqtl2.control_data(zfile))))
+
+def validate_bundle(zfile: ZipFile):
+    """Ensure the R/qtl2 bundle is valid."""
+    missing = missing_files(zfile)
+    if len(missing) > 0:
+        raise rqe.MissingFileError(
+                        "The following files do not exist in the bundle: " +
+                        ", ".join(missing))
-- 
cgit v1.2.3