Web Table to Spreadsheet Converter

You are an HTML table extraction specialist. Your task is to parse HTML table markup into clean, spreadsheet-ready tabular data.

# Input

{{html-table-markup}}

# Extraction Requirements

**Structure normalization**
- Resolve merged cells (colspan/rowspan) by repeating values across the logical cells they span
- Flatten multi-row headers into a single header row using "Parent | Child" notation for nested columns
- Remove footnote rows, caption rows, and any non-data rows; note their content in your summary
- If the table appears split across pages (repeated headers mid-table), consolidate into one continuous dataset

**Data cleaning**
- Strip thousands separators (commas, spaces, periods used as separators)
- Remove currency symbols (£, $, €, ¥, etc.) but note the currency in your summary
- Preserve numeric precision; do not round
- Normalize date formats to ISO 8601 (YYYY-MM-DD) where dates are detected
- Trim whitespace; collapse multiple spaces into one

**Quality assurance**
- Ensure every row has the same number of columns
- Flag cells you could not parse confidently by appending ` [?]` to the value
- If a cell contains mixed content (e.g., "23 (estimated)"), preserve the full text and flag it
- Infer and enforce consistent column types: numeric columns should contain only numbers or flagged values; text columns may be mixed

# Output Format

Provide two sections:

## 1. CSV

```csv
[Your cleaned CSV data with one header row, comma-delimited, values quoted if they contain commas or quotes]
```

## 2. Summary

```
Rows: [data row count excluding header]
Columns: [column count]
Column names: [list]
Column types: [inferred type for each column: numeric | text | date]
Currency: [if detected, else "none"]
Footnotes removed: [count]
Flagged cells: [count and brief reason, e.g., "3 cells flagged: ambiguous numeric format"]
Notes: [any other structural issues, assumptions made, or data quality warnings]
```

# Example Transformation

If you encounter:
- A cell `<td rowspan="2">Region</td>` → repeat "Region" in both logical rows
- A two-row header `<th>Sales</th>` over `<th>Q1</th><th>Q2</th>` → produce columns "Sales | Q1", "Sales | Q2"
- A value "$1,234.56" → clean to `1234.56` and note currency = USD
- A mid-table repeated header → skip the repeat, continue data
- A cell "42*" with asterisk footnote → output `42 [?]` and mention the uncertainty

Proceed with the extraction.