Fixup: PR 811: install an upper-bounded version of Jinja2 during docu…

…mentation build (#812)

* BUG: Fixes issue with Pandas >=1.5

Set objects are no longer allowed as an index argument for a DataFrame
constructor.

Fixes #810

* ENH: Use newer version of node

* (Re)install a semver-upper-bounded version of jinja2 package dependency during sphinx documentation build

Ref: pallets/jinja#1630

Ref: sphinx-doc/sphinx#10291

* Revert "(Re)install a semver-upper-bounded version of jinja2 package dependency during sphinx documentation build"

This reverts commit de26d4d.

* Specify semver-upper-bounded version of jinja2 package-documentation dependency

Follows-on-from commit de26d4d.

* Fixup: flake8-acceptable formatting for setup.py

* Update ChangeLog.md

Co-authored-by: Yoshiki Vázquez Baeza <[email protected]>

.github/workflows/main.yml

@@ -30,7 +30,7 @@ jobs:
       - name: Set up Node.js enviroment
         uses: actions/setup-node@v1
         with:
-          node-version: 14
+          node-version: 16
 
       - uses: conda-incubator/setup-miniconda@v2
         with:

ChangeLog.md

@@ -6,7 +6,7 @@ Emperor ChangeLog
 ### Bug Fixes
 
 * Update remote URL to no longer use rawgit.
-  ([#751(https://github.com/biocore/emperor/issues/751)]).
+  ([#751](https://github.com/biocore/emperor/issues/751)).
 
 ### New Features
 
@@ -16,6 +16,8 @@ Emperor ChangeLog
 
 * Pin Sphinx version to be less than 4.0.
 * Jupyter templates no longer require jQuery to add the CSS headers.
+* Fix broken test suite with Pandas >=1.5 ([#810](https://github.com/biocore/emperor/issues/810)).
+* Limit jinja2 version for doc dependencies in setup.py. Allowed versions are: `>=2.9` and `<3.1`.
 
 # Emperor 1.0.3 (14 Apr 2021)
 -----------------------------

emperor/core.py

@@ -306,9 +306,9 @@ def _validate_metadata(self, metadata, matrix, ignore_missing_samples,
                              'are using metadata and coordinates corresponding'
                              ' to the same dataset.' % kind)
 
+        # sort the elements so we have a deterministic output
+        difference = sorted([str(i) for i in difference])
         if difference and not ignore_missing_samples:
-            # sort the elements so we have a deterministic output
-            difference = sorted([str(i) for i in difference])
 
             # if there's more than 5 missing elements, truncate the list
             if len(difference) > 5:

setup.py

@@ -36,7 +36,8 @@
 
 base = ["numpy >= 1.7", "scipy >= 0.17.0", "click", "pandas",
         "scikit-bio >= 0.4.1", "jinja2 >= 2.9", "future"]
-doc = ["Sphinx<4", "sphinx-bootstrap-theme", "numpydoc"]
+doc = ["Sphinx<4", "jinja2 >= 2.9, < 3.1", "sphinx-bootstrap-theme",
+       "numpydoc"]
 test = ["pep8", "flake8", "nose"]
 all_deps = base + doc + test