Skip to content

Commit

Permalink
Added more filters when querying the MongoDB
Browse files Browse the repository at this point in the history
  • Loading branch information
Ronaldo S.A. Batista committed Nov 9, 2023
1 parent f3cb8b0 commit fe306b9
Show file tree
Hide file tree
Showing 12 changed files with 516 additions and 469 deletions.
10 changes: 1 addition & 9 deletions .gitattributes
Original file line number Diff line number Diff line change
@@ -1,10 +1,2 @@
*.xlsx filter=lfs diff=lfs merge=lfs -text
*.parquet.gzip filter=lfs diff=lfs merge=lfs -text
*.parquet filter=lfs diff=lfs merge=lfs -text
*.pkl filter=lfs diff=lfs merge=lfs -text
*.ipynb merge=nbdev-merge
*.shp filter=lfs diff=lfs merge=lfs -text
*.shx filter=lfs diff=lfs merge=lfs -text
*.prj filter=lfs diff=lfs merge=lfs -text
*.dbf filter=lfs diff=lfs merge=lfs -text
*.cpg filter=lfs diff=lfs merge=lfs -text

2 changes: 1 addition & 1 deletion extracao/__init__.py
Original file line number Diff line number Diff line change
@@ -1 +1 @@
__version__ = "0.8.20"
__version__ = "0.8.21"
29 changes: 12 additions & 17 deletions extracao/anatel.py
Original file line number Diff line number Diff line change
Expand Up @@ -71,8 +71,8 @@ def extraction(self) -> L:
Telecom(self.mongo_uri, self.limit),
SMP(self.mongo_uri, self.limit),
SRD(self.mongo_uri),
# Stel(self.sql_params),
# Radcom(self.sql_params),
Stel(self.sql_params),
Radcom(self.sql_params),
Aero(),
]

Expand Down Expand Up @@ -152,26 +152,20 @@ def intersect_coordinates_on_poligon(
gdf_points.crs = regions.crs

# Spatial join points to the regions
points_with_regions = gpd.sjoin(
gdf_points, regions, how="inner", predicate="within"
)
gdf = gpd.sjoin(gdf_points, regions, how="inner", predicate="within")

if check_municipio:
# Check correctness of Coordinates
check_coords = (
points_with_regions.Código_Município != points_with_regions.CD_MUN
)
check_coords = gdf.Código_Município != gdf.CD_MUN

log = """[("Colunas", ["Código_Município", "Município", "UF"]),
("Processamento", "Informações substituídas pela localização correta das coordenadas.")
"""
self.register_log(points_with_regions, log, check_coords)
self.register_log(gdf, log, check_coords)

points_with_regions.drop(
["Código_Município", "Município", "UF"], axis=1, inplace=True
)
gdf.drop(["Código_Município", "Município", "UF"], axis=1, inplace=True)

points_with_regions.rename(
gdf.rename(
columns={
"CD_MUN": "Código_Município",
"NM_MUN": "Município",
Expand All @@ -180,7 +174,7 @@ def intersect_coordinates_on_poligon(
inplace=True,
)

return points_with_regions
return gdf

def validate_coordinates(
self, df: pd.DataFrame, check_municipio: bool = True
Expand All @@ -207,8 +201,9 @@ def _format(
self,
dfs: List, # List with the individual API sources
) -> pd.DataFrame: # Processed DataFrame
# aero = self.validate_coordinates(dfs.pop(), False)
return self.validate_coordinates(pd.concat(dfs, ignore_index=True))
return pd.concat([aero, anatel], ignore_index=True).sort_values(
aero = self.validate_coordinates(dfs.pop(), False)
anatel = self.validate_coordinates(pd.concat(dfs, ignore_index=True))
df = pd.concat([aero, anatel], ignore_index=True).sort_values(
["Frequência", "Latitude", "Longitude"], ignore_index=True
)
return df.loc[:, self.columns]
8 changes: 7 additions & 1 deletion extracao/constants.py
Original file line number Diff line number Diff line change
Expand Up @@ -82,7 +82,6 @@

AGG_LICENCIAMENTO = [
"Frequência",
"Entidade",
"Fistel",
"Código_Município",
"Longitude",
Expand Down Expand Up @@ -389,6 +388,9 @@
{"NumServico": {"$nin": ["010", "045", "171", "450", "750", "", None]}},
{"FreqTxMHz": {"$nin": [None, "", 0], "$type": 1.0}},
{"CodMunicipio": {"$nin": [None, ""]}},
{"NumFistel": {"$nin": [None, ""]}},
{"CodTipoClasseEstacao": {"$nin": [None, ""]}},
{"DesignacaoEmissao": {"$nin": [None, ""]}},
]
}

Expand All @@ -411,6 +413,10 @@
{"NumServico": "010"},
{"FreqTxMHz": {"$nin": [None, "", 0], "$type": 1.0}},
{"CodMunicipio": {"$nin": [None, ""]}},
{"NumFistel": {"$nin": [None, ""]}},
{"CodTipoClasseEstacao": {"$nin": [None, ""]}},
{"DesignacaoEmissao": {"$nin": [None, ""]}},
{"Tecnologia": {"$nin": [None, ""]}},
]
}

Expand Down
8 changes: 7 additions & 1 deletion nbs/00_constants.ipynb
Original file line number Diff line number Diff line change
Expand Up @@ -134,7 +134,6 @@
"\n",
"AGG_LICENCIAMENTO = [\n",
"\t'Frequência',\n",
"\t'Entidade',\n",
"\t'Fistel',\n",
"\t'Código_Município',\n",
"\t'Longitude',\n",
Expand Down Expand Up @@ -497,6 +496,9 @@
"\t\t{'NumServico': {'$nin': ['010', '045', '171', '450', '750', '', None]}},\n",
"\t\t{'FreqTxMHz': {'$nin': [None, '', 0], '$type': 1.0}},\n",
"\t\t{'CodMunicipio': {'$nin': [None, '']}},\n",
"\t\t{'NumFistel': {'$nin': [None, '']}},\n",
"\t\t{'CodTipoClasseEstacao': {'$nin': [None, '']}},\n",
"\t\t{'DesignacaoEmissao': {'$nin': [None, '']}},\n",
"\t]\n",
"}\n",
"\n",
Expand All @@ -519,6 +521,10 @@
"\t\t{'NumServico': '010'},\n",
"\t\t{'FreqTxMHz': {'$nin': [None, '', 0], '$type': 1.0}},\n",
"\t\t{'CodMunicipio': {'$nin': [None, '']}},\n",
"\t\t{'NumFistel': {'$nin': [None, '']}},\n",
"\t\t{'CodTipoClasseEstacao': {'$nin': [None, '']}},\n",
"\t\t{'DesignacaoEmissao': {'$nin': [None, '']}},\n",
"\t\t{'Tecnologia': {'$nin': [None, '']}},\n",
"\t]\n",
"}"
]
Expand Down
Loading

0 comments on commit fe306b9

Please sign in to comment.