From 99374cb4db0fa47defc39de1facc9212a965a618 Mon Sep 17 00:00:00 2001
From: Johnny Chavez <64660690+calderjo@users.noreply.github.com>
Date: Mon, 25 Nov 2024 15:04:39 -0800
Subject: [PATCH] Fix Kagglehub test and Remove multiple version matplotlib
 files (#1441)

fixing main build.
- kagglehub login changes required updating our tests
- latest matplotlib is needed by handful of package, but breaks
learntools. i provided a workaround
---
 Dockerfile.tmpl         | 39 +++++++++++++++++++++------------------
 tests/test_kagglehub.py | 10 ++++++----
 2 files changed, 27 insertions(+), 22 deletions(-)

diff --git a/Dockerfile.tmpl b/Dockerfile.tmpl
index 11b05ebd..ad6842ea 100644
--- a/Dockerfile.tmpl
+++ b/Dockerfile.tmpl
@@ -167,28 +167,12 @@ RUN pip install jax[cpu] && \
 RUN export PATH=/usr/local/cuda/bin:$PATH && \
     export CUDA_ROOT=/usr/local/cuda && \
     pip install pycuda \
-        pynvrtc \
+        # TODO(379932879): pip resolver fails when not specified.
+        pynvrtc==9.2 \
         pynvml && \
     /tmp/clean-layer.sh
 {{ end }}
 
-# b/308525631: Pin Matplotlib until seaborn can be upgraded
-# to >0.13.0 (now it's stuck by a package conflict with ydata-profiling 4.5.1).
-RUN JAXVER=$(pip freeze | grep -e "^jax==") && \
-    pip install --upgrade \
-        "matplotlib<3.8.0" \
-        # ipympl adds interactive widget support for matplotlib
-        ipympl==0.7.0 \
-        "seaborn==0.12.2" \
-        pyupset \
-        python-dateutil dask dask-expr igraph \
-        pyyaml joblib geopy mne pyshp \
-        pandas \
-        polars \
-        flax \
-        "${JAXVER}" && \
-    /tmp/clean-layer.sh
-
 RUN apt-get update && \
     apt-get install -y default-jre && \
     /tmp/clean-layer.sh
@@ -543,11 +527,30 @@ ENV TESSERACT_PATH=/usr/bin/tesseract \
     # For Theano with MKL
     MKL_THREADING_LAYER=GNU
 
+# b/308525631: Pin Matplotlib until seaborn can be upgraded
+# to >0.13.0 (now it's stuck by a package conflict with ydata-profiling 4.5.1).
+RUN JAXVER=$(pip freeze | grep -e "^jax==") && \
+    pip install --upgrade \
+        "matplotlib==3.7.5" \
+        # ipympl adds interactive widget support for matplotlib
+        ipympl==0.7.0 \
+        "seaborn==0.12.2" \
+        pyupset \
+        python-dateutil dask dask-expr igraph \
+        pyyaml joblib geopy mne pyshp \
+        pandas \
+        polars \
+        flax \
+        "${JAXVER}" && \
+    /tmp/clean-layer.sh
+
 # Temporary fixes and patches
 # Temporary patch for Dask getting downgraded, which breaks Keras
 RUN pip install --upgrade dask && \
     # Stop jupyter nbconvert trying to rewrite its folder hierarchy
     mkdir -p /root/.jupyter && touch /root/.jupyter/jupyter_nbconvert_config.py && touch /root/.jupyter/migrated && \
+    # TODO(b/380921973): Ensure only matplotlib 3.7.5 files are present.
+    rm -r /opt/conda/lib/python3.10/site-packages/matplotlib-3.9.2.dist-info/ && \
     mkdir -p /.jupyter && touch /.jupyter/jupyter_nbconvert_config.py && touch /.jupyter/migrated && \
     # Stop Matplotlib printing junk to the console on first load
     sed -i "s/^.*Matplotlib is building the font cache using fc-list.*$/# Warning removed by Kaggle/g" /opt/conda/lib/python3.10/site-packages/matplotlib/font_manager.py && \
diff --git a/tests/test_kagglehub.py b/tests/test_kagglehub.py
index 37b11248..f2c3e2a6 100644
--- a/tests/test_kagglehub.py
+++ b/tests/test_kagglehub.py
@@ -8,8 +8,10 @@ class TestKagglehub(unittest.TestCase):
     def test_login(self):
         with self.assertLogs('kagglehub', level='INFO') as l:
             with mock.patch("builtins.input") as mock_input:
-                mock_input.side_effect = ["lastplacelarry", "some-key"]
-                # Disabling credentials validation since network access is disabled in unittest.
-                kagglehub.login(validate_credentials=False)
+                with mock.patch("getpass.getpass") as mock_getpass:
+                    mock_input.side_effect = ["lastplacelarry"]
+                    mock_getpass.return_value = "some-key"
 
-                self.assertIn("credentials set", l.output[0])
+                    kagglehub.login(validate_credentials=False)
+
+                    self.assertIn("credentials set", l.output[0])