Ci: test Python download code on PR (#337)

* squash ci/test-downloader

* merge requirements

* rm unused download/requirements

* Run `make clean download import process`

This is an automated update by travis-ci at
Fri Dec 13 00:27:12 PST 2024

[skip ci]

---------

Co-authored-by: OpenDisclosure Deploybot <[email protected]>

Author: ckingbailey

download/.github/workflows/pr_check.yaml

@@ -0,0 +1,32 @@
+name: Python tests
+
+on:
+  pull_request:
+    branches:
+      - feat/pull-v2-api
+  push:
+    branches:
+      - ci/test-downloader
+  workflow_dispatch:
+
+env:
+  working_dir: download
+
+jobs:
+  run_tests:
+    name: Run tests on Python download code
+    runs-on: ubuntu-22.04
+    defaults:
+      run:
+        working-directory: ${{ env.working_dir }}
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v4
+        with:
+          python-version-file: ${{ env.working_dir }}/.python-version
+          cache: pip
+          cache-dependency-path: ${{ env.working_dir }}/requirements.txt
+      - run: pip install -r requirements.txt
+      - name: Run tests
+        run: pytest tests/test_*.py
+  

download/main.py

@@ -32,7 +32,7 @@ def main():
     with open(f'{DATA_DIR_PATH}/filers.json', encoding='utf8') as f:
         filers = json.loads(f.read())
 
-    committees = Committees(filers, elections)
+    committees = Committees(filers, elections.pl)
 
     # A-Contribs:
     # join filers + filings + elections + transactions
@@ -41,13 +41,13 @@ def main():
     #     committees.Ballot_Measure_Election -> elections.Ballot_Measure_Election
     # where trans['transaction']['calTransactionType'] == 'F460A'
     with open(f'{DATA_DIR_PATH}/filings.json', encoding='utf8') as f:
-        filings = Filings(json.loads(f.read()))
+        filings = Filings(json.loads(f.read())).pl
 
     with open(f'{DATA_DIR_PATH}/transactions.json', encoding='utf8') as f:
         records = json.loads(f.read())
-        transactions = Transactions(records)
+        transactions = Transactions(records).pl
 
-    a_contributions = A_Contributions(transactions, filings, committees)
+    a_contributions = A_Contributions(transactions, filings, committees.pl)
     a_contribs_df = a_contributions.df
     if not a_contribs_df.is_empty:
         print(a_contribs_df.drop(columns=[
@@ -72,8 +72,8 @@ def main():
             'XRef_Match',
         ]).sample(n=20))
 
-    elections.df.write_csv(f'{OUTPUT_DIR}/elections.csv')
-    committees.df.write_csv(f'{OUTPUT_DIR}/committees.csv')
+    elections.pl.write_csv(f'{OUTPUT_DIR}/elections.csv')
+    committees.pl.write_csv(f'{OUTPUT_DIR}/committees.csv')
     a_contributions.df.write_csv(f'{OUTPUT_DIR}/a_contributions.csv')
 
 if __name__ == '__main__':

download/model/a_contributions.py

@@ -2,9 +2,7 @@
 Schedule A, Contributions
 Hopefully this can be joined with other Schedule classes into a single Transaction class
 """
-from .committee import Committees
-from .filing import Filings
-from .transaction import Transactions
+import polars as pl
 from .schedule import ScheduleBase
 
 class A_Contributions(ScheduleBase):
@@ -13,9 +11,9 @@ class A_Contributions(ScheduleBase):
     """
     def __init__(
         self,
-        transactions:Transactions,
-        filings:Filings,
-        committees:Committees
+        transactions:pl.DataFrame,
+        filings:pl.DataFrame,
+        committees:pl.DataFrame
     ):
         self._form_id = 'F460A'
         super().__init__(

download/model/base.py

@@ -1,12 +1,13 @@
 """ This is the base model, upon all others shall be based """
+import pandas as pd
 import polars as pl
 
 class BaseModel:
     """ Base model other models inherit from """
     def __init__(self, data):
         self._data = data
         self._df = None
-        self._lazy = None
+        self._pl = None
         self._dtypes = []
         self._pl_dtypes = []
         self._sql_dtypes = []
@@ -19,18 +20,32 @@ def data(self):
         return self._data
 
     @property
-    def lazy(self):
-        ''' Return a Polars Lazyframe '''
-        if self._lazy is None:
-            self._lazy = pl.LazyFrame(self._data, schema=self._dtypes)
-
-        return self._lazy
+    def pl(self):
+        ''' Return a Polars dataframe '''
+        if self._pl is None or self._pl.is_empty():
+            self._pl = pl.DataFrame(self._data, schema=self._pl_dtypes)
 
+        return self._pl
+    
     @property
     def df(self):
-        ''' Return a Polars dataframe '''
-        if self._df is None:
-            self._df = self.lazy.collect()
+        """ Get a dataframe of the data """
+        if self._df is None or self._df.empty:
+            self._df = pd.DataFrame(self._data).astype(self._dtypes)
 
         return self._df
 
+    def to_sql(self, connection, **kwargs):
+        """ Write to a postgresql table """
+        options = {
+            'index_label': 'id',
+            'if_exists': 'replace'
+        }
+        options.update(kwargs)
+
+        self.df[self._sql_cols].to_sql(
+            self._sql_table_name,
+            connection,
+            dtype=self._sql_dtypes,
+            **options
+        )

download/model/committee.py

@@ -2,12 +2,11 @@
 from typing import List
 import polars as pl
 from sqlalchemy.types import String
-# Next line ingored because Pylint reports cannot find election in model
-from . import base, election # pylint: disable=no-name-in-module
+from . import base
 
 class Committees(base.BaseModel):
     """ A collection of committees """
-    def __init__(self, filers:List[dict], elections:election.Elections):
+    def __init__(self, filers:List[dict], elections:pl.DataFrame):
         empty_election_influence = {
             'electionDate': None,
             'measure': None,
@@ -20,6 +19,7 @@ def __init__(self, filers:List[dict], elections:election.Elections):
         super().__init__([
             {
                 'filer_nid': int(f['filerNid']),
+                # 'Ballot_Measure_Election': [ *elections[elections['date'] == infl['electionDate']]['name'].array, None ][0],
                 'Ballot_Measure_Election': self._get_possibly_empty_ballot_measure_election(
                     elections,
                     infl
@@ -49,6 +49,21 @@ def __init__(self, filers:List[dict], elections:election.Elections):
             if f['registrations'].get('CA SOS')
         ])
         self._dtypes = {
+            'filer_nid': int,
+            'Ballot_Measure_Election': 'string',
+            'Filer_ID': 'string',
+            'Filer_NamL': 'string',
+            '_Status': 'string',
+            '_Committee_Type': 'string',
+            'Ballot_Measure': 'string',
+            'Support_Or_Oppose': 'string',
+            'candidate_controlled_id': 'string',
+            'Start_Date': 'string',
+            'End_Date': 'string',
+            'data_warning': 'string',
+            'Make_Active': 'string'
+        }
+        self._pl_dtypes = {
             'filer_nid': pl.UInt64,
             'Ballot_Measure_Election': pl.Utf8,
             'Filer_ID': pl.Utf8,
@@ -63,6 +78,22 @@ def __init__(self, filers:List[dict], elections:election.Elections):
             'data_warning': pl.Utf8,
             'Make_Active': pl.Utf8
         }
+        self._sql_dtypes = {
+            'Ballot_Measure_Election': String,
+            'Filer_ID': String,
+            'Filer_NamL': String,
+            '_Status': String,
+            '_Committee_Type': String,
+            'Ballot_Measure': String,
+            'Support_Or_Oppose': String,
+            'candidate_controlled_id': String,
+            'Start_Date': String,
+            'End_Date': String,
+            'data_warning': String,
+            'Make_Active': String
+        }
+        self._sql_cols = self._sql_dtypes.keys()
+        self._sql_table_name = 'committees'
 
     @staticmethod
     def support_or_oppose(influence):
@@ -85,7 +116,7 @@ def _get_possibly_empty_ballot_measure_election(elections: pl.DataFrame, influen
         list, which will contain either the matched election slug or None.
         '''
         return [
-            *elections.lazy.filter(
+            *elections.lazy().filter(
                 pl.col('date') == influence['electionDate']
             ).first().collect().get_column('name'),
             None

download/model/d_expenditures.py

@@ -1,9 +1,7 @@
 '''
 FPPC Form 460, Schedule D, Expenditures
 '''
-from .committee import Committees
-from .filing import Filings
-from .transaction import Transactions
+import polars as pl
 from .schedule import ScheduleBase
 
 class DExpenditures(ScheduleBase):
@@ -12,9 +10,9 @@ class DExpenditures(ScheduleBase):
     '''
     def __init__(
         self,
-        transactions:Transactions,
-        filings:Filings,
-        committees:Committees
+        transactions: pl.DataFrame,
+        filings: pl.DataFrame,
+        committees: pl.DataFrame
     ):
         self._form_id = 'F460D'
         super().__init__(

download/model/election.py

@@ -3,7 +3,7 @@
 """
 import os
 from datetime import datetime
-from polars import Utf8
+from sqlalchemy.types import String
 from .base import BaseModel
 
 class Elections(BaseModel):
@@ -53,11 +53,19 @@ def __init__(self, election_records):
 
         super().__init__(elections)
         self._dtypes = {
-            'title': Utf8,
-            'name': Utf8,
-            'location': Utf8,
-            'date': Utf8
+            'title': 'string',
+            'name': 'string',
+            'location': 'string',
+            'date': 'string'
         }
+        self._sql_dtypes = {
+            'title': String,
+            'name': String,
+            'location': String,
+            'date': String
+        }
+        self._sql_cols = self._sql_dtypes.keys()
+        self._sql_table_name = 'elections'
 
     @staticmethod
     def ordinal(n):

download/model/filing.py

@@ -17,6 +17,15 @@ def __init__(self, filings):
         ])
 
         self._dtypes = {
+            'filing_nid': 'string',
+            'filer_nid': int,
+            'Report_Num': 'Int64',
+            'Rpt_Date': 'string',
+            'From_Date': 'string',
+            'Thru_Date': 'string'
+        }
+
+        self._pl_dtypes = {
             'filing_nid': Utf8,
             'filer_nid': UInt64,
             'Report_Num': UInt64,

download/model/schedule.py

@@ -2,9 +2,6 @@
 Abstracts much of the boilerplate common to FPPC Form 460 Schedule data
 '''
 import polars as pl
-from .committee import Committees
-from .filing import Filings
-from .transaction import Transactions
 
 DTYPES = {
     'Filer_ID': 'string',
@@ -91,18 +88,18 @@ class ScheduleBase:
     def __init__(
         self,
         form_id: str,
-        transactions:Transactions,
-        filings:Filings,
-        committees:Committees
+        transactions: pl.DataFrame,
+        filings: pl.DataFrame,
+        committees: pl.DataFrame
     ):
-        schedule = committees.df.lazy().group_by('Filer_ID').first().join(
-            filings.df.lazy(),
+        schedule = committees.lazy().group_by('Filer_ID').first().join(
+            filings.lazy(),
             on='filer_nid',
             how='inner'
         ).rename({
             '_Committee_Type': 'Committee_Type'
         }).join(
-            transactions.df.lazy().filter(pl.col('cal_tran_type') == form_id),
+            transactions.lazy().filter(pl.col('cal_tran_type') == form_id),
             on='filing_nid',
             how='inner'
         ).drop([
@@ -121,19 +118,14 @@ def __init__(
         ])
 
         self._lazy = schedule
-        self._df = None
 
         self._dtypes = DTYPES
 
     @property
     def lazy(self):
-        ''' Get data as Polars LazyFrame '''
         return self._lazy
 
     @property
     def df(self):
-        ''' Get data as Polars DataFrame '''
-        if self._df is None:
-            self._df = self._lazy.collect()
-
-        return self._df
+        # QUESTION: Does this invalidate self._lazy?
+        return self._lazy.collect()

download/model/transaction.py

@@ -153,4 +153,4 @@ def __init__(self, transactions):
             } for t in transactions
         ])
 
-        self._dtypes = DTYPES
+        self._pl_dtypes = DTYPES