From 5491da18dda8c6f55bc2bc5d95f21b86908cd382 Mon Sep 17 00:00:00 2001
From: Frederick Muriuki Muriithi
Date: Tue, 9 Jan 2024 09:50:18 +0300
Subject: Raise exception on reading non-existing file

The validation checks ensure that whatever files are listed in the
control file exist in the zip file bundle. It is still possible,
however, that the code tries to read a file that does not exist in the
file and is not listed in the control file. In those cases, raise the
appropriate exception.
---
 r_qtl/r_qtl2.py                             |  35 +++++++++++++++-------------
 tests/r_qtl/test_files/nonexistent.zip      | Bin 0 -> 380 bytes
 tests/r_qtl/test_r_qtl2_nonexistent_file.py |  29 +++++++++++++++++++++++
 3 files changed, 48 insertions(+), 16 deletions(-)
 create mode 100644 tests/r_qtl/test_files/nonexistent.zip
 create mode 100644 tests/r_qtl/test_r_qtl2_nonexistent_file.py

diff --git a/r_qtl/r_qtl2.py b/r_qtl/r_qtl2.py
index a8958a0..d3a3805 100644
--- a/r_qtl/r_qtl2.py
+++ b/r_qtl/r_qtl2.py
@@ -189,23 +189,26 @@ def file_data(zfile: ZipFile,
                   [str, tuple[str, ...], tuple[str, ...]],
                   tuple[dict, ...]] = __default_process_value_transposed__) -> Iterator[dict]:
     """Load data from files in R/qtl2 zip bundle."""
-    if isinstance(cdata[member_key], list):
-        for row in (line for lines in
-                    (file_data(
-                        zfile, member_key, {**cdata, member_key: innerfile},
-                        process_value, process_transposed_value)
-                     for innerfile in cdata[member_key])
-                    for line in lines):
-            yield row
-        return
-    if not cdata.get(f"{member_key}_transposed", False):
-        for row in with_non_transposed(zfile, member_key, cdata, process_value):
-            yield row
-        return
+    try:
+        if isinstance(cdata[member_key], list):
+            for row in (line for lines in
+                        (file_data(
+                            zfile, member_key, {**cdata, member_key: innerfile},
+                            process_value, process_transposed_value)
+                         for innerfile in cdata[member_key])
+                        for line in lines):
+                yield row
+            return
+        if not cdata.get(f"{member_key}_transposed", False):
+            for row in with_non_transposed(zfile, member_key, cdata, process_value):
+                yield row
+            return
 
-    for row in with_transposed(
-            zfile, member_key, cdata, process_transposed_value):
-        yield row
+        for row in with_transposed(
+                zfile, member_key, cdata, process_transposed_value):
+            yield row
+    except KeyError as exc:
+        raise InvalidFormat(*exc.args) from exc
 
 def cross_information(zfile: ZipFile, cdata: dict) -> Iterator[dict]:
     """Load cross information where present."""
diff --git a/tests/r_qtl/test_files/nonexistent.zip b/tests/r_qtl/test_files/nonexistent.zip
new file mode 100644
index 0000000..4a31bf1
Binary files /dev/null and b/tests/r_qtl/test_files/nonexistent.zip differ
diff --git a/tests/r_qtl/test_r_qtl2_nonexistent_file.py b/tests/r_qtl/test_r_qtl2_nonexistent_file.py
new file mode 100644
index 0000000..c783a76
--- /dev/null
+++ b/tests/r_qtl/test_r_qtl2_nonexistent_file.py
@@ -0,0 +1,29 @@
+"""Test accessing nonexistent member."""
+from pathlib import Path
+
+import pytest
+from zipfile import ZipFile
+
+from r_qtl import r_qtl2 as rqtl2
+
+@pytest.mark.unit_test
+@pytest.mark.parametrize(
+    "filepath,memberkey",
+    (("tests/r_qtl/test_files/nonexistent.zip", "geno"),
+     ("tests/r_qtl/test_files/nonexistent.zip", "founder_geno"),
+     ("tests/r_qtl/test_files/nonexistent.zip", "pheno"),
+     ("tests/r_qtl/test_files/nonexistent.zip", "covar"),
+     ("tests/r_qtl/test_files/nonexistent.zip", "phenocovar"),
+     ("tests/r_qtl/test_files/nonexistent.zip", "gmap"),
+     ("tests/r_qtl/test_files/nonexistent.zip", "pmap")))
+def test_loading_nonexistent_file(filepath, memberkey):
+    """
+    GIVEN: A zipfile with a valid control file, but some files mentioned in the
+        control file do not exist in the zipfile
+    WHEN: access is made to such missing files
+    THEN: raise an exception
+    """
+    with (ZipFile(Path(filepath).absolute(), "r") as zfile,
+          pytest.raises(rqtl2.InvalidFormat)):
+        cdata = rqtl2.control_data(zfile)
+        tuple(rqtl2.file_data(zfile, memberkey, cdata))
-- 
cgit v1.2.3