collector/plugins/finance/robor_plugin.py

83 lines
2.4 KiB
Python
Raw Normal View History

2020-04-15 20:29:29 +00:00
import datetime
import lxml.html
import requests
from peewee import *
from playhouse.shortcuts import model_to_dict
import config
from database import BaseModel
from plugins.plugin import Plugin
class Robor(BaseModel):
date = DateField(index=True, default=datetime.date.today(), null=False)
field = TextField(null=False)
value = FloatField(null=False)
class RoborPlugin(Plugin):
models = [Robor]
def __init__(self):
self.__table = None
2020-04-17 16:16:58 +00:00
def get_interval(self):
return config.ROBOR_INTERVAL
2020-04-15 20:29:29 +00:00
def get_column_index(self, table, column_name):
header_row = table.find('tr')
for elem in header_row.iter('th'):
if column_name in elem.text_content():
return header_row.index(elem)
return None
def values_newer_than(self, table, oldest_date: datetime.date, col_date, col_value):
# Items are ordered descending, so stop when we reach date
for row in table.iter('tr'):
dt_str = row[col_date].text_content()
val_str = row[col_value].text_content()
try:
dt = datetime.datetime.strptime(dt_str, '%d/%b/%Y')
val = float(val_str)
except ValueError:
continue
dt = dt.date()
if (oldest_date is not None) and (dt <= oldest_date):
break
yield (dt, val)
def execute(self):
# Get last existing date
latest_date = Robor.select(Robor.date) \
.order_by(Robor.date.desc()) \
.limit(1) \
.scalar()
# Fetch & parse data
response = requests.get('https://www.bnro.ro/StatisticsReportHTML.aspx?icid=801&table=642', verify=False)
response.raise_for_status()
html = lxml.html.fromstring(response.text)
# Read data from table
table = html.find('.//table[@class="stat_table"]')
for field in config.ROBOR_FIELDS:
col_date = self.get_column_index(table, 'Date')
col_value = self.get_column_index(table, field)
for date, value in self.values_newer_than(table, latest_date, col_date, col_value):
entry = Robor()
entry.date = date
entry.field = field
entry.value = value
entry.save()