Exported PI invoices as PDFs

The PI-specific dataframes will first be converted to HTML
tables using Jinja templates, and then converted to PDFs using
the Selenium Chrome driver.

A html template folder has been added, and the test cases
for the PI-specific invoice will now both check whether the
dataframe is formatted correctly and if the PDFs are
correctly generated.

process_report/invoices/pi_specific_invoice.py

@@ -1,12 +1,21 @@
 import os
 from dataclasses import dataclass
+import base64
+import tempfile
 
 import pandas
+from jinja2 import Environment, FileSystemLoader
+from selenium import webdriver
+from selenium.webdriver.common.by import By
+from selenium.webdriver.common.print_page_options import PrintOptions
 
 import process_report.invoices.invoice as invoice
 import process_report.util as util
 
 
+TEMPLATE_DIR_PATH = "process_report/templates"
+
+
 @dataclass
 class PIInvoice(invoice.Invoice):
     """
@@ -15,6 +24,21 @@ class PIInvoice(invoice.Invoice):
     - NewPICreditProcessor
     """
 
+    TOTAL_COLUMN_LIST = [
+        invoice.COST_FIELD,
+        invoice.CREDIT_FIELD,
+        invoice.BALANCE_FIELD,
+    ]
+
+    DOLLAR_COLUMN_LIST = [
+        invoice.RATE_FIELD,
+        invoice.GROUP_BALANCE_FIELD,
+        invoice.COST_FIELD,
+        invoice.GROUP_BALANCE_USED_FIELD,
+        invoice.CREDIT_FIELD,
+        invoice.BALANCE_FIELD,
+    ]
+
     export_columns_list = [
         invoice.INVOICE_DATE_FIELD,
         invoice.PROJECT_FIELD,
@@ -37,30 +61,134 @@ class PIInvoice(invoice.Invoice):
         invoice.BALANCE_FIELD,
     ]
 
+    @staticmethod
+    def _get_chrome_driver():
+        options = webdriver.ChromeOptions()
+        options.add_argument("--headless")
+        return webdriver.Chrome(options=options)
+
     def _prepare(self):
         self.data = self.data[
             self.data[invoice.IS_BILLABLE_FIELD] & ~self.data[invoice.MISSING_PI_FIELD]
         ]
         self.pi_list = self.data[invoice.PI_FIELD].unique()
 
+    def _get_pi_dataframe(self, data, pi):
+        def add_dollar_sign(data):
+            if pandas.isna(data):
+                return data
+            else:
+                return "$" + str(data)
+
+        pi_projects = data[data[invoice.PI_FIELD] == pi].copy().reset_index(drop=True)
+
+        # Remove prepay group data if it's empty
+        if pandas.isna(pi_projects[invoice.GROUP_NAME_FIELD]).all():
+            pi_projects = pi_projects.drop(
+                [
+                    invoice.GROUP_NAME_FIELD,
+                    invoice.GROUP_INSTITUTION_FIELD,
+                    invoice.GROUP_BALANCE_FIELD,
+                    invoice.GROUP_BALANCE_USED_FIELD,
+                ],
+                axis=1,
+            )
+
+        # Add a row containing sums for certain columns
+        column_sums = list()
+        sum_columns_list = list()
+        for column_name in self.TOTAL_COLUMN_LIST:
+            if column_name in pi_projects.columns:
+                column_sums.append(pi_projects[column_name].sum())
+                sum_columns_list.append(column_name)
+        pi_projects.loc[
+            len(pi_projects)
+        ] = None  # Adds a new row to end of dataframe initialized with None
+        pi_projects.loc[pi_projects.index[-1], invoice.INVOICE_DATE_FIELD] = "Total"
+        pi_projects.loc[pi_projects.index[-1], sum_columns_list] = column_sums
+
+        # Add dollar sign to certain columns
+        for column_name in self.DOLLAR_COLUMN_LIST:
+            if column_name in pi_projects.columns:
+                pi_projects[column_name] = pi_projects[column_name].apply(
+                    add_dollar_sign
+                )
+
+        pi_projects.fillna("", inplace=True)
+
+        return pi_projects
+
     def export(self):
-        def _export_pi_invoice(pi):
-            if pandas.isna(pi):
-                return
-            pi_projects = export_data[export_data[invoice.PI_FIELD] == pi]
-            pi_instituition = pi_projects[invoice.INSTITUTION_FIELD].iat[0]
-            pi_projects.to_csv(
-                f"{self.name}/{pi_instituition}_{pi} {self.invoice_month}.csv"
+        def _create_html_invoice():
+            environment = Environment(loader=FileSystemLoader(TEMPLATE_DIR_PATH))
+            template = environment.get_template("pi_invoice.html")
+            content = template.render(
+                data=pi_dataframe,
             )
 
+            with tempfile.NamedTemporaryFile(
+                mode="w", delete=False, suffix=".html"
+            ) as f:
+                f.write(content)
+                return f.name
+
+        def _create_pdf_invoice():
+            driver = self._get_chrome_driver()
+            driver.get("file://" + invoice_html_path)
+
+            """
+            To ensure the HTML invoice table always fit the page when
+            printed to PDF, certain things are done below. Before my
+            explanation, some context is needed:
+
+            When printing to pdf, the Chrome driver picks the default
+            page size of A4, which is 8.27in x 11.69in. In CSS, the length
+            units include pixels and inches, and it seems 1 inch
+            equals 96 pixels. This explains the constants below.
+
+            In order to ensure the invoice table fits along the width
+            (since we’re printing landscape) of the A4 page, we want
+            to scale the table down.
+
+            This scaling value should be =  width of A4 / length of table.
+            """
+            A4_WIDTH_INCHES = 11.69
+            PX_PER_INCH = 96
+            invoice_table = driver.find_element(by=By.TAG_NAME, value="table")
+            px_width = float(
+                invoice_table.value_of_css_property("width").split("px")[0]
+            )
+            scale_value = A4_WIDTH_INCHES / (px_width / PX_PER_INCH)
+
+            print_options = PrintOptions()
+            print_options.orientation = "landscape"
+            print_options.scale = scale_value
+            print_options.shrink_to_fit = True
+
+            pdf_str = driver.print_page(print_options=print_options)
+            pdf = base64.b64decode(pdf_str)
+            with open(
+                f"{self.name}/{pi_instituition}_{pi}_{self.invoice_month}.pdf", "wb"
+            ) as f:
+                f.write(pdf)
+
         export_data = self._filter_columns()
+
         if not os.path.exists(
             self.name
         ):  # self.name is name of folder storing invoices
             os.mkdir(self.name)
 
         for pi in self.pi_list:
-            _export_pi_invoice(pi)
+            if pandas.isna(pi):
+                continue
+
+            pi_dataframe = self._get_pi_dataframe(export_data, pi)
+            pi_instituition = pi_dataframe[invoice.INSTITUTION_FIELD].iat[0]
+
+            invoice_html_path = _create_html_invoice()
+            _create_pdf_invoice()
+            os.remove(invoice_html_path)
 
     def export_s3(self, s3_bucket):
         def _export_s3_pi_invoice(pi_invoice):

process_report/templates/pi_invoice.html

@@ -0,0 +1,59 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="utf-8">
+  </head>
+
+  <style>
+      body {
+        -webkit-print-color-adjust: exact;
+        print-color-adjust: exact;
+      }
+      table {
+        font-family: arial, sans-serif;
+        border-collapse: collapse;
+        width: 100%;
+      }
+      td, th {
+        border: 1px solid #8d8d8d;
+        text-align: left;
+        padding: 8px;
+      }
+      th {
+        text-align: center;
+      }
+      tr:nth-child(even) {
+        background-color: #dddddd;
+      }
+      tr:last-child {
+        background-color: #dddddd;
+        font-weight: bold;
+      }
+  </style>
+
+  <body>
+    <table>
+      <tr>
+          {% for col in data.columns %}
+            <th>{{col}}</th>
+          {% endfor %}
+      </tr>
+
+      {% for i, row in data.iterrows() %}
+        <tr>
+          {% for field in row %}
+            {% if i == data.index[-1] %}
+              {% if field %}
+                <th>{{field}}</th>
+              {% else %}
+                <td style="border-width: 0;"></td>
+              {% endif %}
+            {% else %}
+              <td>{{field}}</td>
+            {% endif %}
+          {% endfor %}
+        </tr>
+      {% endfor %}
+    </table>
+  </body>
+</html>