def LoadFromMongo(self,
find_filter,
db_name="AnalysisEvo",
coll_name="Simulation",
client="local"):
MS = MongoStorage(client=client)
meta_list = [
j for j in MS.FilterAndGetMetadata(
find_filter=find_filter, db_name=db_name, coll_name=coll_name)
]
if len(meta_list) == 0:
raise ValueError(
"Could not find any simulation series from the given find_filter!"
)
original_seed_list = [
j.get("extra").get("original_seed") for j in meta_list
]
id_list = [j.get("_id") for j in meta_list]
meta_df = pd.DataFrame(
dict(original_seed=original_seed_list,
_id=id_list)).sort_values("original_seed")
self.result_rs_list = [
MS.LoadAllSeriesFromId([j], db_name=db_name,
coll_name=coll_name)[0]
for j in meta_df["_id"].values
]
self.startdate = meta_list[0].get("extra").get("startdate")
self.enddate = meta_list[0].get("extra").get("enddate")
def SampleStrategy(self, random_n_sample = None, filter_dict = {}, dont_load = False):
if filter_dict is None:
filter_dict = {"User": {"$nin": ["Deleted"]}, "Name": {"$regex": "_[0-9]{4}_"}}
MS = MongoStorage()
meta_list = [j for j in MS.FilterAndGetMetadata(filter_dict)]
df = pd.DataFrame(meta_list)
df["filter1"] = df.groupby("Name")["SavedDate"].apply(lambda x: x==np.max(x)) # conflicting
df["filter2"] = df.groupby("Name")["EndDate"].apply(lambda x: x==np.max(x))
df["filter"] = df["filter2"] #& df["filter1"]
fdf = df[df["filter"]].sort_values("Name")
fdf = fdf.set_index("Name")
fdf = fdf.loc[~fdf.index.duplicated(keep='last')]
fdf = fdf.reset_index()
if random_n_sample is not None:
fdf = fdf.sample(random_n_sample)
id_list = fdf["_id"].values
name_list = fdf["Name"].values
if dont_load:
return name_list
else:
status = [self.AddSeries(j) for j in MS.LoadAllSeriesFromId(id_list)]