compose and style data dictionary HTML

NYCPlanning · Dec 2, 2024 · 9b365ce · 9b365ce
1 parent 65b2420
commit 9b365ce
Show file tree

Hide file tree

Showing 13 changed files with 105 additions and 113 deletions.
diff --git a/dcpy/lifecycle/package/generate_metadata_assets.py b/dcpy/lifecycle/package/generate_metadata_assets.py
@@ -1,5 +1,5 @@
 from pathlib import Path
-from jinja2 import Template, Environment, FileSystemLoader
+from jinja2 import Environment, FileSystemLoader
 import css_inline
 from bs4 import BeautifulSoup
 import subprocess
@@ -9,9 +9,12 @@
 from . import RESOURCES_PATH
 
 DEFAULT_DATA_DICTIONARY_TEMPLATE_PATH = (
-    RESOURCES_PATH / "data_dictionary_template.jinja"
+    RESOURCES_PATH / "document_templates" / "data_dictionary.jinja"
 )
-DEFAULT_DATA_DICTIONARY_STYLESHEET_PATH = RESOURCES_PATH / "data_dictionary.css"
+DEFAULT_DATA_DICTIONARY_STYLESHEET_PATH = (
+    RESOURCES_PATH / "document_templates" / "document.css"
+)
+DEFAULT_PDF_STYLESHEET_PATH = RESOURCES_PATH / "document_templates" / "paged_media.css"
 
 
 def _format_html(html: str) -> str:
@@ -48,10 +51,9 @@ def _style_html_document(html: str, stylesheet_path: Path) -> str:
 def generate_pdf_from_html(
     html_path: Path,
     output_path: Path,
-    stylesheet_path: Path,
+    stylesheet_path: Path = DEFAULT_PDF_STYLESHEET_PATH,
 ) -> Path:
     logger.info(f"Saving DCP PDF to {output_path}")
-    # TODO style HTML before converting to PDF
     subprocess.run(
         [
             "weasyprint",
@@ -69,17 +71,16 @@ def generate_html_from_yaml(
     metadata_path: Path,
     output_path: Path,
     html_template_path: Path,
+    stylesheet_path: Path,
 ) -> Path:
-    metadata = Metadata.from_path(metadata_path, template_vars={"var1": "value1"})
+    metadata = Metadata.from_path(metadata_path)
 
-    # TODO use _render_html_template and _compose_html_document
-    with open(html_template_path, "r") as f:
-        template_text = f.read()
-    rendered_template = Template(template_text).render({"metadata": metadata})
+    html = _render_html_template(html_template_path, {"metadata": metadata})
+    styled_html = _style_html_document(html, stylesheet_path)
 
     output_path.parent.mkdir(parents=True, exist_ok=True)
     logger.info(f"Saving DCP HTML to {output_path}")
     with open(output_path, "w") as f:
-        f.write(rendered_template)
+        f.write(styled_html)
 
     return output_path
diff --git a/dcpy/lifecycle/package/resources/DCP_Logo.svg b/dcpy/lifecycle/package/resources/DCP_Logo.svg
diff --git a/dcpy/lifecycle/package/resources/data_dictionary.css b/dcpy/lifecycle/package/resources/data_dictionary.css
diff --git a/dcpy/lifecycle/package/resources/data_dictionary_template.jinja b/dcpy/lifecycle/package/resources/data_dictionary_template.jinja
diff --git a/dcpy/lifecycle/package/resources/dcp_logo.png b/dcpy/lifecycle/package/resources/dcp_logo.png
diff --git a/dcpy/lifecycle/package/resources/document_templates/attributes.jinja b/dcpy/lifecycle/package/resources/document_templates/attributes.jinja
@@ -0,0 +1,21 @@
+<div class="attributes">
+
+  <h2>Attributes</h2>
+
+  <div class="field">
+
+    <p class="field-content">
+      <span class="field-name">Description:</span> {{ metadata.attributes.description }}
+    </p>
+
+    <p class="field-content">
+      <span class="field-name">Each row is a:</span> {{ metadata.attributes.each_row_is_a }}
+    </p>
+
+    <p class="field-content">
+      <span class="field-name">Tags:</span> {{ metadata.attributes.tags|join(', ') }}
+    </p>
+
+  </div>
+
+</div>
diff --git a/dcpy/lifecycle/package/resources/document_templates/columns.jinja b/dcpy/lifecycle/package/resources/document_templates/columns.jinja
@@ -0,0 +1,28 @@
+<div class="attributes">
+
+  <h2>Columns</h2>
+
+  {% for column in metadata.columns %}
+  <div class="column">
+
+    <h3>{{ column.name }}</h3>
+
+    <p class="column-field">
+      <span class="field-name">Description:</span> {{ column.description }}
+    </p>
+    <p class="column-field">
+      <span class="field-name">Type:</span>
+      {% if column.custom.readme_data_type %}
+      {{ column.custom.readme_data_type }}
+      {% else %}
+      {{ column.data_type }}
+      {% endif %}
+    </p>
+    <p class="column-field">
+      <span class="field-name">Source:</span> {{ column.data_source }}
+    </p>
+
+  </div>
+  {% endfor %}
+
+</div>
diff --git a/dcpy/lifecycle/package/resources/document_templates/data_dictionary.jinja b/dcpy/lifecycle/package/resources/document_templates/data_dictionary.jinja
@@ -0,0 +1,17 @@
+<!DOCTYPE html>
+<html lang="en">
+
+<head>
+    <meta charset="UTF-8">
+    <title>{{ metadata.attributes.display_name }} Data Dictionary</title>
+</head>
+
+<body>
+    <h1>{{ metadata.attributes.display_name }} - Data Dictionary</h1>
+
+    {% include 'attributes.jinja' %}
+
+    {% include 'columns.jinja' %}
+</body>
+
+</html>
diff --git a/dcpy/lifecycle/package/resources/document_templates/document.css b/dcpy/lifecycle/package/resources/document_templates/document.css
@@ -0,0 +1,12 @@
+html {
+  color: #393939;
+  font-family: Fira Sans;
+  font-size: 10pt;
+}
+
+.field-name {
+  font-weight: bold;
+}
+.column-field {
+  text-indent: 1cm;
+}
diff --git a/dcpy/lifecycle/package/resources/document_templates/paged_media.css b/dcpy/lifecycle/package/resources/document_templates/paged_media.css
@@ -0,0 +1,12 @@
+@page {
+  @top-left {
+    content: url("../dcp_logo.png");
+  }
+  @top-right {
+    content: string(top-right-text);
+  }
+  @bottom-right-corner {
+    content: counter(page);
+    text-align: center;
+  }
+}
diff --git a/dcpy/test/lifecycle/package/test_generate_data_dictionary.py b/dcpy/test/lifecycle/package/test_generate_data_dictionary.py
@@ -77,6 +77,7 @@ def test_generate_pdf_from_yaml(self):
             metadata_path=self.yaml_path,
             output_path=self.html_path,
             html_template_path=generate_metadata_assets.DEFAULT_DATA_DICTIONARY_TEMPLATE_PATH,
+            stylesheet_path=generate_metadata_assets.DEFAULT_DATA_DICTIONARY_STYLESHEET_PATH,
         )
         pdf_path = generate_metadata_assets.generate_pdf_from_html(
             html_path=html_path,

diff --git a/products/template/build_scripts/export.py b/products/template/build_scripts/export.py
@@ -31,11 +31,11 @@ def generate_metadata():
         dataset_metadata_yml,
         PRODUCT_PATH / "data_dictionary.html",
         generate_metadata_assets.DEFAULT_DATA_DICTIONARY_TEMPLATE_PATH,
+        generate_metadata_assets.DEFAULT_DATA_DICTIONARY_STYLESHEET_PATH,
     )
     generate_metadata_assets.generate_pdf_from_html(
         html_path,
         PRODUCT_PATH / "data_dictionary.pdf",
-        generate_metadata_assets.DEFAULT_DATA_DICTIONARY_STYLESHEET_PATH,
     )
     oti_xlsx.write_oti_xlsx(
         dataset=md.Metadata.from_path(dataset_metadata_yml).dataset,

diff --git a/pyproject.toml b/pyproject.toml
@@ -150,5 +150,5 @@ library = "dcpy.library.cli:run"
 [tool.setuptools.package-data]
 dcpy = [
     "library/templates/*.yml",
-    "lifecycle/package/resources/*",
+    "lifecycle/package/resources/**",
 ]