Fix inverted resolution check in _line_to_pnts (under-resolved line warning)

[martinguthrie93]

Changes proposed in this PR:

• Fix the inverted resolution check in climada.util.lines_polys_handler._line_to_pnts. The mask used to count under-resolved lines selected lines longer than 10 * resolution (line_lengths > 10 * res), while the inline comment and the log message both describe lines shorter than 10 * resolution. The comparison is now < 10 * res, so the warning is raised for the lines it is meant to flag.
• Correct the wording and typos of the warning message (disaggregate → disaggregated, Reaggregatint → Reaggregating, overestimattion → overestimation, chosing → choosing).
• Update test_resolution_warning, which previously asserted the buggy behavior (it expected 2 flagged lines for input lengths [2, 12, 20] with res = 1, even though lengths 12 and 20 are longer than 10 * res = 10). The expected count is now 1, corresponding to the single genuinely under-resolved line (length 2).

Why this matters

When a line exposure (road, pipeline, transmission line, river reach, …) is disaggregated into points, the number of points is round(length / res). A line that is short relative to res collapses to very few points — in the limit, a single representative point — and re-aggregating an impact computed on too few points typically overestimates that feature's impact. The warning exists to alert users to this situation.

Because the comparison was inverted:

• users with genuinely under-resolved lines received no warning (silent overestimation), and
• users with well-resolved long lines received a warning whose text contradicted itself.

Only the warning path is affected; the returned geometry is unchanged.

Verification

The existing test test_resolution_warning had encoded the contradiction (asserting "2 lines with a length < 10*resolution" for two lines that are in fact longer than 10 * res). I reproduced the mask logic in isolation to confirm the corrected behavior:

import numpy as np
line_lengths = np.array([2.0, 12.0, 20.0]); res = 1  # 10 * res == 10
len(line_lengths[line_lengths > 10 * res])  # old: 2 (the LONG lines 12, 20)
len(line_lengths[line_lengths < 10 * res])  # new: 1 (the short line 2)

Note: the full CLIMADA test suite was not run in my local environment (CLIMADA and its geospatial dependencies were not installed there); the change is confined to the warning branch and its dedicated unit test, which I verified matches exactly.

This PR fixes #1302

PR Author Checklist

• Read the Contribution Guide
• Correct target branch selected (if unsure, select develop)
• Descriptive pull request title added
• Source branch up-to-date with target branch
• Documentation updated
• Tests updated
• Tests passing (verified the affected unit test's logic in isolation; the full suite was not run locally — see note above)
• No new linter issues (black reports both changed files unchanged)
• Changelog updated