The Problem
We have a raw list of 3,000 root domains (e.g., `example-site.com`). However, we need to find specific articles or sub-pages (e.g., `example-site.com/specific-topic-article`) that are relevant to our client's niche.

The solution must be capable of taking a homepage as input, "expanding" the site structure to find deep content, and filtering out irrelevant pages (e.g., "About Us", generic news, or off-topic content).

Input Data
You are provided with two datasets:
- 'Input_Domains.csv`: A raw export of 3,000 root domains. Note: This is "messy" real-world data containing both high-relevance sites and irrelevant "junk" sites.
Target_Keywords.csv: A list of specific keywords that define "Relevance" for this project.

The Task
Your solution should process the inputs and execute the following pipeline:
1.  Expansion & Crawling: Iterate through the domains and crawl the site to discover relevant deep pages (articles, guides, reviews).
2.  Semantic Filtering: Score each identified page against the `Target_Keywords.csv` list. A page is a "Match" if it contains target keywords in its core metadata (Title, H1, URL).
3. Exclusion: Automatically discard "utility" pages (e.g., Privacy Policy, Login) and irrelevant topics.

Deliverables
1.  Source Code / Methodology: The complete code, script, or workflow used to process the list.
2.  Final Spreadsheet (CSV): The output file generated by your solution using our provided sample data.

Success Metrics
- Precision: Does the output successfully exclude the irrelevant domains included in the input?
- Granularity: Does the output provide specific deep URLs rather than just homepages?