feat: improve Python notebook cell output (#1505)

[stu-code]

Summary:

Improvements to Python code blocks based on feature request 1505:

• Suppress SAS log for Python cells; show output only when there is Python stdout, a Python error, or a SAS-level error:

• Render Python exceptions using VS Code's built-in error renderer with coloring and highlighting

• Python code that fails to run results in an X on the notebook cell

• New default - Auto-display the last expression in a Python cell: objects with repr_html() (DataFrames, sklearn models, pipelines) are rendered as rich HTML via base64 extraction; all other values fall back to print(). A second optional mode, SAS.show(), renders via ODS HTML5 and is configurable via SAS.notebook.python.autoPrintMethod. Configuration is available in UI.

Testing:
Tests were performed manually in a Python code block with the exception of (4) which uses a SAS code block to create a SAS error condition. (See TODOs for updating build tests)

1. Tests for if there is no Python stdout or an error:

Test 1.1:

import pandas as pd

Expected result: No output

Test 1.2:

foo = 'bar'

Expected result: No output

Test 1.3:

SAS.symput('foo', 'bar')

Expected result: No output

---

2. Tests for Python output:

Test 2.1:

40+2

Expected result:

42

Test 2.2:

{'Hello': 'world'}

Expected result:

{'Hello': 'world'}

Test 2.3:

asdf

Expected result:

--------------------------------------------------------------------------------------------------------------------------------

NameError                                 Traceback (most recent call last)
  File "<stdin>", line 5, in <module>
  File "<stdin>", line 2, in <module>
  File "<string>", line 1, in <module>
NameError: name 'asdf' is not defined

---

3. Test DataFrame and model rendering.

Test 3.1:

import pandas as pd
df = pd.DataFrame({'foo': [1, 2, 3], 'bar': [4, 5, 6]})
df

from sklearn.ensemble import GradientBoostingClassifier
model = GradientBoostingClassifier()
model

Test 3.2: Switch to SAS.show from (auto) in SAS extension Python autoprint settings:

import pandas as pd
df = pd.DataFrame({'foo': [1, 2, 3], 'bar': [4, 5, 6]})
df

from sklearn.ensemble import GradientBoostingClassifier
model = GradientBoostingClassifier()
model

Expected result: DataFrame renders in ODS HTML5. Model produces the following:

This method doesn't know the object passed in. This is an object of type <class 'sklearn.ensemble._gb.GradientBoostingClassifier'>.
GradientBoostingClassifier()

---

4. Test SAS errors: run the following SAS code, then run a Python code block:

Test 4.1:

options set=PROC_PYPATH="/foo";

proc python restart;
run;

'Hello world!'

Expected result:

ERROR: Failed to create Python subprocess.
ERROR: Launch Failed. Cannot execute program "/foo", error = 2 (No such file or directory).
ERROR: The specified executable module either cannot be located or it is not a valid executable.
ERROR: Failed to launch Python sub-process in tkpy extension.

TODOs:

• Update build test cases for new expected Python output behavior
• Translate new Python autoprint option to other languages

[stu-code]

Note that in order for this to pass the build test, it will need to be updated for the new Python code block output behavior.