diff --git a/service/service.py b/service/service.py index c5bf64f..aa9bb4d 100644 --- a/service/service.py +++ b/service/service.py @@ -165,7 +165,6 @@ class PointService: point.age_day = AGE_DAY_LIMIT point.save() - # print(qs.dist.m) @staticmethod def get_min_distances_to_group(postamat_id: str): diff --git a/service/tasks.py b/service/tasks.py index 3a76eed..74a1f27 100644 --- a/service/tasks.py +++ b/service/tasks.py @@ -27,7 +27,7 @@ from service.layer_service import LayerService @shared_task() -def raschet(): +def raschet(table_name='service_placementpoint'): LastMLCall.objects.all().delete() LastMLCall.objects.create() log_to_telegram('start raschet') @@ -46,23 +46,17 @@ def raschet(): pts['geometry'] = pts['geometry'].apply(wkb.loads, hex=True) pts = gpd.GeoDataFrame(pts, geometry='geometry', crs='epsg:4326') pts = pts.to_crs('epsg:32637') - pts = pts.rename( - columns={ - 'target_cnt_nearby_mean': 'target_dist1', - 'target_age_nearby_mean': 'target_dist2', - 'yndxfood_cnt_cst': 'target_dist3', - }, - ) + feats = [ 'id', 'metro_dist', 'target_dist', 'property_price_bargains', 'property_price_offers', 'property_mean_floor', - 'property_era', 'flats_cnt_2', 'flats_cnt', 'popul_home', 'popul_job', 'other_post_cnt', 'yndxfood_sum', + 'property_era', 'flats_cnt', 'popul_home', 'popul_job', 'yndxfood_sum', 'yndxfood_cnt', 'school_cnt', 'kindergar_cnt', 'target_post_cnt', 'public_stop_cnt', 'sport_center_cnt', 'pharmacy_cnt', 'supermarket_cnt', 'supermarket_premium_cnt', 'clinic_cnt', 'bank_cnt', 'reca_cnt', 'lab_cnt', 'culture_cnt', 'attraction_cnt', 'mfc_cnt', 'bc_cnt', 'tc_cnt', 'rival_pvz_cnt', 'rival_post_cnt', - 'business_activity', 'age_day', 'target_cnt_ao_mean', 'target_dist1', 'target_dist2', 'target_dist3', + 'business_activity', 'age_day', 'target_cnt_ao_mean' ] # Записи для обучения @@ -83,23 +77,6 @@ def raschet(): ) pts_trn.loc[pts_trn.target_dist > 700, 'target_dist'] = 700 - pts_trn['target_dist1'] = pts_trn.apply( - lambda x: ((sorted(distance.cdist([[x['geometry'].x, x['geometry'].y]], target_feature_coords)[0])[2])), - axis=1, - ) - pts_trn.loc[pts_trn.target_dist1 > 700, 'target_dist1'] = 700 - - pts_trn['target_dist2'] = pts_trn.apply( - lambda x: ((sorted(distance.cdist([[x['geometry'].x, x['geometry'].y]], target_feature_coords)[0])[3])), - axis=1, - ) - pts_trn.loc[pts_trn.target_dist2 > 700, 'target_dist2'] = 700 - - pts_trn['target_dist3'] = pts_trn.apply( - lambda x: ((sorted(distance.cdist([[x['geometry'].x, x['geometry'].y]], target_feature_coords)[0])[4])), - axis=1, - ) - pts_trn.loc[pts_trn.target_dist3 > 700, 'target_dist3'] = 700 pts_trn['buf'] = pts_trn.buffer(500) pts_trn = gpd.GeoDataFrame(pts_trn, geometry='buf', crs='epsg:32637') @@ -113,11 +90,17 @@ def raschet(): Y_trn = pts_trn[['fact']] # Записи для инференса - pts_inf = pts.loc[(pts.status == 'Pending') | - (pts.status == 'Installation') | - (pts.status == 'Cancelled') | - ((pts.status == 'Working') & (pts.sample_trn == False))].reset_index(drop=True) - pts_inf = gpd.GeoDataFrame(pts_inf, geometry='geometry', crs='epsg:32637') + if table_name == 'service_placementpoint': + pts_inf = pts.loc[(pts.status == 'Pending') | + (pts.status == 'Installation') | + (pts.status == 'Cancelled') | + ((pts.status == 'Working') & (pts.sample_trn == False))].reset_index(drop=True) + elif table_name == 'service_preplacementpoint': + pts_inf = pd.read_sql(f"select * from {table_name}", conn) + pts_inf = pts_inf.loc[pts_inf.matching_status == 'New'].reset_index(drop=True) + pts_inf['geometry'] = pts_inf['geometry'].apply(wkb.loads, hex=True) + pts_inf = gpd.GeoDataFrame(pts_inf, geometry='geometry', crs='epsg:4326') + pts_inf = pts_inf.to_crs('epsg:32637') pts_inf['buf'] = pts_inf.buffer(500) pts_inf = gpd.GeoDataFrame(pts_inf, geometry='buf', crs='epsg:32637') @@ -139,24 +122,6 @@ def raschet(): ) pts_inf.loc[pts_inf.target_dist > 700, 'target_dist'] = 700 - pts_inf['target_dist1'] = pts_inf.apply( - lambda x: ((sorted(distance.cdist([[x['geometry'].x, x['geometry'].y]], target_feature_coords)[0])[1])), - axis=1, - ) - pts_inf.loc[pts_inf.target_dist1 > 700, 'target_dist1'] = 700 - - pts_inf['target_dist2'] = pts_inf.apply( - lambda x: ((sorted(distance.cdist([[x['geometry'].x, x['geometry'].y]], target_feature_coords)[0])[2])), - axis=1, - ) - pts_inf.loc[pts_inf.target_dist2 > 700, 'target_dist2'] = 700 - - pts_inf['target_dist3'] = pts_inf.apply( - lambda x: ((sorted(distance.cdist([[x['geometry'].x, x['geometry'].y]], target_feature_coords)[0])[3])), - axis=1, - ) - pts_inf.loc[pts_inf.target_dist3 > 700, 'target_dist3'] = 700 - pts_inf = pts_inf.sort_values(by='id').reset_index(drop=True) target_post = gpd.sjoin(pts_inf, pts_target, op='contains').groupby('id', as_index=False).agg({'cnt': 'count'}) target_post = target_post.rename(columns={'cnt': 'target_post_cnt'}) @@ -251,7 +216,7 @@ def raschet(): update_records1 = [] for i in range(0, len(update_fields)): update_records1.append((int(update_fields.prediction_current[i]), int(update_fields.id[i]))) - sql_update_query = """Update service_placementpoint set prediction_current = %s where id = %s""" + sql_update_query = f"""Update {table_name} set prediction_current = %s where id = %s""" try: psycopg2.extras.execute_batch(cursor, sql_update_query, update_records1) conn2.commit() @@ -264,7 +229,7 @@ def raschet(): update_records2 = [] for i in range(0, len(update_fields_working)): update_records2.append((int(update_fields_working.plan_first[i]), int(update_fields_working.id[i]))) - sql_update_query = """Update service_placementpoint set plan_first = %s where id = %s""" + sql_update_query = f"""Update {table_name} set plan_first = %s where id = %s""" try: psycopg2.extras.execute_batch(cursor, sql_update_query, update_records2) conn2.commit() @@ -277,7 +242,7 @@ def raschet(): update_records3 = [] for i in range(0, len(update_fields_working)): update_records3.append((int(update_fields_working.plan_current[i]), int(update_fields_working.id[i]))) - sql_update_query = """Update service_placementpoint set plan_current = %s where id = %s""" + sql_update_query = f"""Update {table_name} set plan_current = %s where id = %s""" try: psycopg2.extras.execute_batch(cursor, sql_update_query, update_records3) conn2.commit() @@ -290,7 +255,7 @@ def raschet(): update_records4 = [] for i in range(0, len(update_fields_working)): update_records4.append((int(update_fields_working.delta_first[i]), int(update_fields_working.id[i]))) - sql_update_query = """Update service_placementpoint set delta_first = %s where id = %s""" + sql_update_query = f"""Update {table_name} set delta_first = %s where id = %s""" try: psycopg2.extras.execute_batch(cursor, sql_update_query, update_records4) conn2.commit() @@ -303,7 +268,7 @@ def raschet(): update_records5 = [] for i in range(0, len(update_fields_working)): update_records5.append((int(update_fields_working.delta_current[i]), int(update_fields_working.id[i]))) - sql_update_query = """Update service_placementpoint set delta_current = %s where id = %s""" + sql_update_query = f"""Update {table_name} set delta_current = %s where id = %s""" try: psycopg2.extras.execute_batch(cursor, sql_update_query, update_records5) conn2.commit() @@ -316,6 +281,7 @@ def raschet(): log_to_telegram('end raschet') + @shared_task def load_post_and_pvz(obj_id: int): file = models.TempFiles.objects.get(id=obj_id) diff --git a/service/views.py b/service/views.py index 728fe7e..c851d30 100644 --- a/service/views.py +++ b/service/views.py @@ -363,6 +363,7 @@ class PrePlacementPointViewSet(PlacementPointViewSet): file_id = request.POST['id'] total, matched, problem = PointService().start_mathing(file_id) PointService.make_enrichment() + raschet('service_preplacementpoint') return Response( {'message': {'total': total, 'matched': matched, 'error': problem, 'unmatched': total - matched - problem}}, status=HTTPStatus.OK,