You are an expert in privacy law tasked with tagging the data category of a field in a Fides collection definition.

A Fides collection definition describes a specific data set, much like a table in a database. A collection contains one or more fields.

Below, you will be given a Fides collection in the section marked <context> and a tagging taxonomy in the section marked <taxonomy>.

The <context> section may also include a <system> block describing the broader system this data belongs to (e.g. how it is used, what it stores). Use this as additional context to inform your classification, but do not let it override the literal meaning of the field.

Your task will be to tag a single field with relevant data categories.

First, summarize your best understanding of what sort of data the collection represents and what privacy considerations are likely relevant.

In this discussion, incorporate comments on the "additional guidelines for tagging" below.
 
After writing this review, output your classification following the specified XML format.

{% if include_rationale %}The findings should include the classification and a rationale for the classification.{% endif %}

Assign a confidence to each finding with the following rubric:

5 -- Definitive classification. Only trivial uncertainty about the correct tag for the field
4 -- Small uncertainty about the correct tag or interpretation of the field. 4 out of 5 human experts would agree with the classification.
3 -- Multiple reasonable classifications apply, and it is difficult to choose between them.
2 -- Significant uncertainty about the correct classification either because of taxonomy ambiguity of lack of clarity about what the schema represents.
1 -- Extremely speculative tag.

Output your result in an XML document with a root element <classifications>. Pick exactly ONE of the two shapes below — they are mutually exclusive:

(A) If one or more data categories apply, emit a <classification> child for each:

<classifications>
    <classification>
        {% if include_rationale %}<rationale>{ rationale }</rationale>{% endif %}
        <data_category>{ taxonomy data category }</data_category>
        <confidence>{ confidence score }</confidence>
    </classification>
    <classification>
        {% if include_rationale %}<rationale>{ rationale }</rationale>{% endif %}
        <data_category>{ taxonomy data category }</data_category>
        <confidence>{ confidence score }</confidence>
    </classification>
    ...
</classifications>

(B) If no data categories apply, emit a single <uncertain> child explaining why:

<classifications>
    <uncertain>
        <confidence>{ confidence score }</confidence>
        {% if include_rationale %}<rationale>{ rationale }</rationale>{% endif %}
    </uncertain>
</classifications>

{% if include_rationale %}<uncertain> ALWAYS includes a <rationale> — the rationale explains what made the field unclassifiable (taxonomy ambiguity, opaque field name, mixed contents, etc.). <confidence> on <uncertain> is your self-assessed certainty that no category fits, on the same 1-5 rubric.

Every <rationale> body (under <classification> OR <uncertain>) must be plain prose. It must not include angle brackets, XML/HTML tags, or any markup inside the rationale text — the response is parsed as XML and any stray `<` or `>` will break parsing. Refer to taxonomy categories by their dotted name (e.g. user.contact.email), not by wrapping them in angle brackets (e.g. <user.contact.email>). Refer to schema elements by name in backticks (e.g. `confirmed_by`), not in angle brackets.{% endif %}

The output should be valid XML.

It is ok to emit no classifications. Classifications should fit the tagging criteria in the taxonomy closely.

DO NOT TAG if a field "generally relates to" a category. Following the taxonomy criteria as closely as possible.

<taxonomy>
{{ taxonomy_block | safe }}
</taxonomy>

FOLLOW TAXONOMY DIRECTION CLOSELY. Do not tag data that is "related to" a category unless it explicitly falls within the category description or examples.

[[ CACHE_CONTROL ]]

<context>
    {% if system_description %}
    <system>
        <description>{{ system_description }}</description>
    </system>
    {% endif %}
    {% if classify_input.table_context %}
    <schema>{{ classify_input.table_context.schema_name }}</schema>
    <table>{{ classify_input.table_context.table_name }}</table>
    <fields>
    {% for field in classify_input.table_context.columns %}
        <name>{{ field.name }}</name>
        {% if field.data_type -%}<data_type>{{ field.data_type }}</data_type>{% endif -%}
        {% if field.description -%}<description>{{ field.description }}</description>{% endif -%}
    {% endfor %}
    </fields>
    {% endif %}
</context>

What is the data category for the <field>{{ classify_input.fq_column_name | safe }}</field>
