junte/Model
1
0
Fork 0
mirror of https://github.com/Hestia-Homes/Model.git synced 2026-06-08 11:17:27 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

added bulk

Browse source
This commit is contained in:
Jun-te Kim 2026-03-31 15:51:43 +00:00
parent 1ac5cb253a
commit ea84cf9fd4
6 changed files with 236 additions and 16 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

24
backend/app/db/models/organisation.py
View file

@ -40,6 +40,30 @@ class HubspotDealData(SQLModel, table=True):
coordination_status: Optional[str] = Field(default=None)
design_status: Optional[str] = Field(default=None)
listing_id: Optional[str] = Field(default=None)
pashub_link: Optional[str] = Field(default=None)
sharepoint_link: Optional[str] = Field(default=None)
dampmould_growth: Optional[str] = Field(default=None)
pre_sap: Optional[str] = Field(default=None)
coordinator: Optional[str] = Field(default=None)
mtp_completion_date: Optional[datetime] = Field(default=None)
mtp_re_model_completion_date: Optional[datetime] = Field(default=None)
ioe_v3_completion_date: Optional[datetime] = Field(default=None)
proposed_measures: Optional[str] = Field(default=None)
approved_package: Optional[str] = Field(default=None)
designer: Optional[str] = Field(default=None)
design_completion_date: Optional[datetime] = Field(default=None)
actual_measures_installed: Optional[str] = Field(default=None)
installer: Optional[str] = Field(default=None)
installer_handover: Optional[str] = Field(default=None)
lodgement_status: Optional[str] = Field(default=None)
measures_lodgement_date: Optional[datetime] = Field(default=None)
lodgement_date: Optional[datetime] = Field(default=None)
expected_commencement_date: Optional[datetime] = Field(default=None)
surveyor: Optional[str] = Field(default=None)
confirmed_survey_date: Optional[datetime] = Field(default=None)
confirmed_survey_time: Optional[str] = Field(default=None)
created_at: datetime = Field(
sa_column=Column(
DateTime(timezone=True),

31
etl/hubspot/hubspotClient.py
View file

@ -189,6 +189,7 @@ class HubspotClient:
)
listing_info: dict[str, str] = cast(dict[str, str], listing.properties) # type: ignore[reportUnknownMemberType]
listing_info["listing_id"] = listing_id
self.logger.info(f"Listing info for deal {deal_id}: {listing_info}")
return listing_info
@ -201,13 +202,35 @@ class HubspotClient:
"dealname",
"dealstage",
"pipeline",
"outcome", # outcome,
"outcome_notes", # outcome notes
"outcome",
"outcome_notes",
"project_code",
"major_condition_issue_description",
"major_condition_issue_photos",
"coordination_status__stage_1_", # Coordiantion Status (Stage 1),
"retrofit_design_status", # Retrofit Design Status
"coordination_status__stage_1_",
"retrofit_design_status",
"pashub_link",
"sharepoint_link",
"dampmould_growth",
"pre_sap",
"coordinator",
"mtp_completion_date",
"mtp_re_model_completion_date",
"ioe_v3_completion_date",
"proposed_measures",
"approved_package",
"designer",
"design_completion_date",
"actual_measures_installed",
"installer",
"installer_handover",
"lodgement_status",
"measures_lodgement_date",
"lodgement_date",
"expected_commencement_date",
"surveyor",
"confirmed_survey_date",
"confirmed_survey_time",
],
)

154
etl/hubspot/hubspotDataTodB.py
View file

@ -2,7 +2,7 @@ from backend.app.db.connection import db_read_session
from backend.app.db.models.organisation import Organisation, HubspotDealData
from sqlmodel import select
from datetime import datetime, timezone
from typing import TypedDict
from typing import TypedDict, Optional
from etl.hubspot.s3_uploader import S3Uploader
import hashlib
import os
@ -82,6 +82,14 @@ class HubspotDataToDb:
.one_or_none()
)
def _parse_hs_date(self, value: Optional[str]) -> Optional[datetime]:
if not value:
return None
try:
return datetime.fromisoformat(value.replace("Z", "+00:00"))
except ValueError:
return None
def _sha256(self, file_path: str) -> str:
"""Compute SHA-256 checksum of a file."""
sha256 = hashlib.sha256()
@ -114,6 +122,10 @@ class HubspotDataToDb:
deal_in_db.deal_id == hs_deal.get("hs_object_id"), "deal_id mismatch"
),
soft_assert(deal_in_db.company_id == hs_company_id, "company_id mismatch"),
soft_assert(
deal_in_db.listing_id == hs_listing.get("listing_id"),
"listing_id mismatch",
),
soft_assert(
deal_in_db.landlord_property_id == hs_listing.get("owner_property_id"),
"landlord_property_id mismatch",
@ -157,6 +169,94 @@ class HubspotDataToDb:
deal_in_db.design_status == hs_deal.get("retrofit_design_status"),
"retrofit design mismatch",
),
soft_assert(
deal_in_db.pashub_link == hs_deal.get("pashub_link"),
"pashub_link mismatch",
),
soft_assert(
deal_in_db.sharepoint_link == hs_deal.get("sharepoint_link"),
"sharepoint_link mismatch",
),
soft_assert(
deal_in_db.dampmould_growth == hs_deal.get("dampmould_growth"),
"dampmould_growth mismatch",
),
soft_assert(
deal_in_db.pre_sap == hs_deal.get("pre_sap"),
"pre_sap mismatch",
),
soft_assert(
deal_in_db.coordinator == hs_deal.get("coordinator"),
"coordinator mismatch",
),
soft_assert(
deal_in_db.mtp_completion_date == self._parse_hs_date(hs_deal.get("mtp_completion_date")),
"mtp_completion_date mismatch",
),
soft_assert(
deal_in_db.mtp_re_model_completion_date == self._parse_hs_date(hs_deal.get("mtp_re_model_completion_date")),
"mtp_re_model_completion_date mismatch",
),
soft_assert(
deal_in_db.ioe_v3_completion_date == self._parse_hs_date(hs_deal.get("ioe_v3_completion_date")),
"ioe_v3_completion_date mismatch",
),
soft_assert(
deal_in_db.proposed_measures == hs_deal.get("proposed_measures"),
"proposed_measures mismatch",
),
soft_assert(
deal_in_db.approved_package == hs_deal.get("approved_package"),
"approved_package mismatch",
),
soft_assert(
deal_in_db.designer == hs_deal.get("designer"),
"designer mismatch",
),
soft_assert(
deal_in_db.design_completion_date == self._parse_hs_date(hs_deal.get("design_completion_date")),
"design_completion_date mismatch",
),
soft_assert(
deal_in_db.actual_measures_installed == hs_deal.get("actual_measures_installed"),
"actual_measures_installed mismatch",
),
soft_assert(
deal_in_db.installer == hs_deal.get("installer"),
"installer mismatch",
),
soft_assert(
deal_in_db.installer_handover == hs_deal.get("installer_handover"),
"installer_handover mismatch",
),
soft_assert(
deal_in_db.lodgement_status == hs_deal.get("lodgement_status"),
"lodgement_status mismatch",
),
soft_assert(
deal_in_db.measures_lodgement_date == self._parse_hs_date(hs_deal.get("measures_lodgement_date")),
"measures_lodgement_date mismatch",
),
soft_assert(
deal_in_db.lodgement_date == self._parse_hs_date(hs_deal.get("lodgement_date")),
"lodgement_date mismatch",
),
soft_assert(
deal_in_db.expected_commencement_date == self._parse_hs_date(hs_deal.get("expected_commencement_date")),
"expected_commencement_date mismatch",
),
soft_assert(
deal_in_db.surveyor == hs_deal.get("surveyor"),
"surveyor mismatch",
),
soft_assert(
deal_in_db.confirmed_survey_date == self._parse_hs_date(hs_deal.get("confirmed_survey_date")),
"confirmed_survey_date mismatch",
),
soft_assert(
deal_in_db.confirmed_survey_time == hs_deal.get("confirmed_survey_time"),
"confirmed_survey_time mismatch",
),
]
# If discrepancies found, update from HubSpot
@ -238,6 +338,7 @@ class HubspotDataToDb:
for attr, value in {
"dealname": deal_data.get("dealname"),
"dealstage": deal_data.get("dealstage"),
"listing_id": listing.get("listing_id"),
"landlord_property_id": listing.get("owner_property_id"),
"uprn": listing.get("national_uprn"),
"outcome": deal_data.get("outcome"),
@ -250,16 +351,32 @@ class HubspotDataToDb:
"major_condition_issue_photos": deal_data.get(
"major_condition_issue_photos"
),
"major_condition_issue_description": deal_data.get(
"major_condition_issue_description"
),
"major_condition_issue_photos": deal_data.get(
"major_condition_issue_photos"
),
"coordination_status": deal_data.get(
"coordination_status__stage_1_"
),
"design_status": deal_data.get("retrofit_design_status"),
"pashub_link": deal_data.get("pashub_link"),
"sharepoint_link": deal_data.get("sharepoint_link"),
"dampmould_growth": deal_data.get("dampmould_growth"),
"pre_sap": deal_data.get("pre_sap"),
"coordinator": deal_data.get("coordinator"),
"mtp_completion_date": self._parse_hs_date(deal_data.get("mtp_completion_date")),
"mtp_re_model_completion_date": self._parse_hs_date(deal_data.get("mtp_re_model_completion_date")),
"ioe_v3_completion_date": self._parse_hs_date(deal_data.get("ioe_v3_completion_date")),
"proposed_measures": deal_data.get("proposed_measures"),
"approved_package": deal_data.get("approved_package"),
"designer": deal_data.get("designer"),
"design_completion_date": self._parse_hs_date(deal_data.get("design_completion_date")),
"actual_measures_installed": deal_data.get("actual_measures_installed"),
"installer": deal_data.get("installer"),
"installer_handover": deal_data.get("installer_handover"),
"lodgement_status": deal_data.get("lodgement_status"),
"measures_lodgement_date": self._parse_hs_date(deal_data.get("measures_lodgement_date")),
"lodgement_date": self._parse_hs_date(deal_data.get("lodgement_date")),
"expected_commencement_date": self._parse_hs_date(deal_data.get("expected_commencement_date")),
"surveyor": deal_data.get("surveyor"),
"confirmed_survey_date": self._parse_hs_date(deal_data.get("confirmed_survey_date")),
"confirmed_survey_time": deal_data.get("confirmed_survey_time"),
}.items():
setattr(existing, attr, value or getattr(existing, attr))
@ -302,6 +419,7 @@ class HubspotDataToDb:
deal_id=deal_id,
dealname=deal_data.get("dealname"),
dealstage=deal_data.get("dealstage"),
listing_id=listing.get("listing_id"),
landlord_property_id=listing.get("owner_property_id"),
uprn=listing.get("national_uprn"),
outcome=deal_data.get("outcome"),
@ -316,6 +434,28 @@ class HubspotDataToDb:
),
coordination_status=deal_data.get("coordination_status__stage_1_"),
design_status=deal_data.get("retrofit_design_status"),
pashub_link=deal_data.get("pashub_link"),
sharepoint_link=deal_data.get("sharepoint_link"),
dampmould_growth=deal_data.get("dampmould_growth"),
pre_sap=deal_data.get("pre_sap"),
coordinator=deal_data.get("coordinator"),
mtp_completion_date=self._parse_hs_date(deal_data.get("mtp_completion_date")),
mtp_re_model_completion_date=self._parse_hs_date(deal_data.get("mtp_re_model_completion_date")),
ioe_v3_completion_date=self._parse_hs_date(deal_data.get("ioe_v3_completion_date")),
proposed_measures=deal_data.get("proposed_measures"),
approved_package=deal_data.get("approved_package"),
designer=deal_data.get("designer"),
design_completion_date=self._parse_hs_date(deal_data.get("design_completion_date")),
actual_measures_installed=deal_data.get("actual_measures_installed"),
installer=deal_data.get("installer"),
installer_handover=deal_data.get("installer_handover"),
lodgement_status=deal_data.get("lodgement_status"),
measures_lodgement_date=self._parse_hs_date(deal_data.get("measures_lodgement_date")),
lodgement_date=self._parse_hs_date(deal_data.get("lodgement_date")),
expected_commencement_date=self._parse_hs_date(deal_data.get("expected_commencement_date")),
surveyor=deal_data.get("surveyor"),
confirmed_survey_date=self._parse_hs_date(deal_data.get("confirmed_survey_date")),
confirmed_survey_time=deal_data.get("confirmed_survey_time"),
)
# Handle upload at insert time

33
etl/hubspot/scripts/scraper/bulk_load.py Normal file
View file

@ -0,0 +1,33 @@
from etl.hubspot.hubspotClient import HubspotClient, Companies, Pipeline
from etl.hubspot.scripts.scraper.main import handler
from tqdm import tqdm
PIPELINE_ID = Pipeline.OPERATIONS_SOCIAL_HOUSING.value
companies = list([Companies.THE_GUINESS_PARTNERSHIP, Companies.SOUTHERN_HOUSING_GROUP])
def bulk_load(companies: list[Companies] | None = None) -> None:
"""
Load all deals from the given companies (defaults to all Companies enum values)
into the database, filtered to the Operations/Social Housing pipeline.
"""
hubspot = HubspotClient()
targets = companies or list(Companies)
for company in tqdm(targets, desc="Companies"):
company_id = company.value
deal_ids = hubspot.get_deal_ids_from_company(company_id)
for deal_id in tqdm(deal_ids, desc=f"{company.name}", leave=False):
deal_data = hubspot.from_deal_id_get_info(deal_id)
if deal_data.get("pipeline") != PIPELINE_ID:
continue
handler({"hubspot_deal_id": deal_id}, context=None)
print(f"Processed deal {deal_id} (company: {company.name})")
if __name__ == "__main__":
bulk_load(companies)

2
etl/hubspot/scripts/scraper/main.py
View file

@ -4,7 +4,7 @@
3) [completed] Load the db and check if upsert it into the table
4) [completed]Getting working on a AWS lambda
5) [completed] subtask and tasks history
6) [TODO]The new sexy deal properties, move it over
6) [completed]The new sexy deal properties, move it over
"""
from etl.hubspot.hubspotClient import HubspotClient

8
sfr/principal_pitch/2_export_data.py
View file

@ -26,13 +26,13 @@ from backend.app.db.functions.materials_functions import get_materials
from collections import defaultdict
from sqlalchemy import func
PORTFOLIO_ID = 639
SCENARIOS = [1157]
PORTFOLIO_ID = 640
SCENARIOS = [1154]
scenario_names = {
1157: "EPC C - no EWI solid floor",
1154: "EPC - 10k Budget",
}
project_name = "Instagroup Sample"
project_name = "First Charterhouse Investments"
def get_data(portfolio_id, scenario_ids):