Eight source-side defects in python_by_example, functions, numpy and pandas (found via translation review)

[mmcky]

Eight defects surfaced by GitHub Copilot's automatic review of the Malayalam translation PRs in QuantEcon/lecture-python-programming.ml (PRs 1–5). Each was checked against this repo at 5816589 and is present verbatim in the English source — the translations copy the source faithfully, so these are source-side issues rather than translation issues. Line numbers are for 5816589.

#	File	Line(s)	Issue
1	lectures/python_by_example.md	383, 385	Inline code spans are written with triple backticks (```i < ts_length```) inside a paragraph. Triple backticks are fence syntax; inline spans should use single backticks. Four occurrences on these two lines
2	lectures/functions.md	454	A stack is described as "a First In Last Out (FILO) queue". The standard term is Last In, First Out (LIFO)
3	lectures/numpy.md	218	`np.loadtxt` ---see [the documentation] — no space before the em-dash marker
4	lectures/numpy.md	1409, 1413	"Numpy" should be "NumPy"; **Part1** is missing a space, and the matching solution heading at 1472 reads **Part 1 Solution**
5	lectures/numpy.md	1274	Prose references an instance of `discreteRV` but the class defined at 1246 is DiscreteRV
6	lectures/pandas.md	254	df.query("cc + cg >= 80 & POP <= 20000") combines comparisons with & without parentheses. The preceding boolean-indexing example uses explicit parentheses, and & binds tighter than the comparisons, so the two examples are not equivalent as the surrounding prose implies
7	lectures/pandas.md	539	Prose refers to https://research.stlouisfed.org/fred2/series/UNRATE/downloaddata/UNRATE.csv, but the surrounding code cells use https://fred.stlouisfed.org/graph/fredgraph.csv?.... The research.stlouisfed.org host no longer serves that path
8	lectures/pandas.md	383, 406	type(x)!=str does not treat str subclasses as strings; isinstance(x, str) is the idiomatic check. At 406 the following np.isnan(x) can also raise on non-float object values, where pd.isna(x) would not

Items 1–5 are typography and naming; 6–8 are code-correctness. Items 6 and 8 change behaviour and are worth fixing first.

Reported from the Malayalam edition's calibration review; the same defects will be present in every translated edition of these lectures, so fixing them here propagates through the normal sync.