Skip to content

Commit

Permalink
partial disagg accessblty
Browse files Browse the repository at this point in the history
  • Loading branch information
@jpn--
jpn-- committed Nov 11, 2023
1 parent a5f1476 commit 7bce5f2
Show file tree
Hide file tree
Showing 2 changed files with 101 additions and 57 deletions.
148 changes: 95 additions & 53 deletions activitysim/abm/models/disaggregate_accessibility.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@
import logging
import random
from functools import reduce
from typing import Any, Literal

import numpy as np
import pandas as pd
Expand All @@ -14,36 +15,87 @@
from activitysim.abm.models.util import tour_destination
from activitysim.abm.tables import shadow_pricing
from activitysim.core import estimation, los, tracing, util, workflow
from activitysim.core.configuration.base import PreprocessorSettings, PydanticReadable
from activitysim.core.configuration.logit import TourLocationComponentSettings
from activitysim.core.expressions import assign_columns

logger = logging.getLogger(__name__)


def read_disaggregate_accessibility_yaml(state: workflow.State, file_name):
class DisaggregateAccessibilitySuffixes(PydanticReadable):
SUFFIX: str = "proto_"
ROOTS: list[str] = [
"persons",
"households",
"tours",
"persons_merged",
"person_id",
"household_id",
"tour_id",
]


class DisaggregateAccessibilityTableSettings(PydanticReadable, extra="forbid"):
index_col: str | None = None
zone_col: str | None = None
rename_columns: dict[str, str] = {}
VARIABLES: dict[str, int | list[int]]
mapped_fields: dict[str, dict]


class DisaggregateAccessibilityAnnotateSettings(PydanticReadable, extra="forbid"):
tablename: str
annotate: PreprocessorSettings


class DisaggregateAccessibilitySettings(PydanticReadable, extra="forbid"):
suffixes: DisaggregateAccessibilitySuffixes = {"SUFFIX": None, "ROOTS": []}
ORIGIN_SAMPLE_SIZE: int | float = 0
DESTINATION_SAMPLE_SIZE: int | float = 0
BASE_RANDOM_SEED: int = 0
add_size_tables: bool = True
zone_id_names: dict[str, str] = {}
ORIGIN_SAMPLE_METHOD: Literal[
None, "full", "uniform", "uniform-taz", "kmeans"
] = None
ORIGIN_WEIGHTING_COLUMN: str
CREATE_TABLES: dict[str, Any] = {}
MERGE_ON: dict[str, list[str]]
FROM_TEMPLATES: bool = False
annotate_proto_tables: list[DisaggregateAccessibilityAnnotateSettings]
NEAREST_METHOD: str = "skims"


def read_disaggregate_accessibility_yaml(
state: workflow.State, file_name
) -> DisaggregateAccessibilitySettings:
"""
Adds in default table suffixes 'proto_' if not defined in the settings file
"""
model_settings = state.filesystem.read_model_settings(file_name)
if not model_settings.get("suffixes"):
model_settings["suffixes"] = {
"SUFFIX": "proto_",
"ROOTS": [
"persons",
"households",
"tours",
"persons_merged",
"person_id",
"household_id",
"tour_id",
],
}
model_settings = DisaggregateAccessibilitySettings.read_settings_file(
state.filesystem, file_name
)
# if not model_settings.get("suffixes"):
# model_settings["suffixes"] = {
# "SUFFIX": "proto_",
# "ROOTS": [
# "persons",
# "households",
# "tours",
# "persons_merged",
# "person_id",
# "household_id",
# "tour_id",
# ],
# }
# Convert decimal sample rate to integer sample size
for sample in ["ORIGIN_SAMPLE_SIZE", "DESTINATION_SAMPLE_SIZE"]:
size = model_settings.get(sample, 0)
size = getattr(model_settings, sample)
if size > 0 and size < 1:
model_settings[sample] = round(
size * len(state.get_dataframe("land_use").index)
setattr(
model_settings,
sample,
round(size * len(state.get_dataframe("land_use").index)),
)

return model_settings
Expand All @@ -66,19 +118,16 @@ def __init__(self, state: workflow.State, network_los, chunk_size):
)

# Random seed
self.seed = self.model_settings.get("BASE_RANDOM_SEED", 0) + len(
self.land_use.index
)
self.seed = self.model_settings.BASE_RANDOM_SEED + len(self.land_use.index)

# Generation
self.params = self.read_table_settings()
self.create_proto_pop()
logger.info(
"Created a proto-population with %s households across %s origin zones to %s possible destination zones"
% (
"Created a proto-population with {} households across {} origin zones to {} possible destination zones".format(
len(self.proto_pop["proto_households"]),
len(self.proto_pop["proto_households"].home_zone_id.unique()),
self.model_settings["DESTINATION_SAMPLE_SIZE"],
self.model_settings.DESTINATION_SAMPLE_SIZE,
)
)
self.inject_tables(state)
Expand All @@ -88,11 +137,11 @@ def __init__(self, state: workflow.State, network_los, chunk_size):
# - initialize shadow_pricing size tables after annotating household and person tables
# since these are scaled to model size, they have to be created while single-process
# this can now be called as a standalone model step instead, add_size_tables
add_size_tables = self.model_settings.get("add_size_tables", True)
add_size_tables = self.model_settings.add_size_tables
if add_size_tables:
# warnings.warn(f"Calling add_size_tables from initialize will be removed in the future.", FutureWarning)
shadow_pricing.add_size_tables(
state, self.model_settings.get("suffixes"), scale=False
state, self.model_settings.suffixes.dict(), scale=False
)

def zone_sampler(self):
Expand Down Expand Up @@ -131,14 +180,14 @@ def zone_sampler(self):

# default_zone_col = 'TAZ' if not (self.network_los.zone_system == los.ONE_ZONE) else 'zone_id'
# zone_cols = self.model_settings["zone_id_names"].get("zone_group_cols", default_zone_col)
id_col = self.model_settings["zone_id_names"].get("index_col", "zone_id")
method = self.model_settings.get("ORIGIN_SAMPLE_METHOD")
n_samples = self.model_settings.get("ORIGIN_SAMPLE_SIZE", 0)
id_col = self.model_settings.zone_id_names.get("index_col", "zone_id")
method = self.model_settings.ORIGIN_SAMPLE_METHOD
n_samples = self.model_settings.ORIGIN_SAMPLE_SIZE

# Get weights, need to get households first to get persons merged.
# Note: This will cause empty zones to be excluded. Which is intended, but just know that.
zone_weights = self.land_use[
self.model_settings["ORIGIN_WEIGHTING_COLUMN"]
self.model_settings.ORIGIN_WEIGHTING_COLUMN
].to_frame("weight")
zone_weights = zone_weights[zone_weights.weight != 0]

Expand Down Expand Up @@ -177,12 +226,12 @@ def zone_sampler(self):
# Join the land_use pop on centroids,
# this also filter only zones we need (relevant if running scaled model)
centroids_df = centroids_df.join(
self.land_use[self.model_settings["ORIGIN_WEIGHTING_COLUMN"]],
self.land_use[self.model_settings.ORIGIN_WEIGHTING_COLUMN],
how="inner",
)
xy_list = list(centroids_df[["X", "Y"]].itertuples(index=False, name=None))
xy_weights = np.array(
centroids_df[self.model_settings["ORIGIN_WEIGHTING_COLUMN"]]
centroids_df[self.model_settings.ORIGIN_WEIGHTING_COLUMN]
)

# Initializer k-means class
Expand Down Expand Up @@ -303,14 +352,9 @@ def zone_sampler(self):
def read_table_settings(self):
# Check if setup properly

assert "CREATE_TABLES" in self.model_settings.keys()

# Set zone_id name if not already specified
self.model_settings["zone_id_names"] = self.model_settings.get(
"zone_id_names", {"index_col": "zone_id"}
)
create_tables = self.model_settings.get("CREATE_TABLES")
from_templates = self.model_settings.get("FROM_TEMPLATES", False)
create_tables = self.model_settings.CREATE_TABLES
from_templates = self.model_settings.FROM_TEMPLATES
zone_list = self.zone_sampler()
params = {}

Expand Down Expand Up @@ -458,14 +502,14 @@ def create_proto_pop(self):
klist = ["proto_households", "proto_persons", "proto_tours"]

# Create ID columns, defaults to "%tablename%_id"
hhid, perid, tourid = [
hhid, perid, tourid = (
self.params[x]["index_col"]
if len(self.params[x]["index_col"]) > 0
else x + "_id"
for x in klist
]
)

if self.model_settings.get("FROM_TEMPLATES"):
if self.model_settings.FROM_TEMPLATES:
table_params = {k: self.params.get(k) for k in klist}
tables = {
k: pd.read_csv(
Expand All @@ -476,7 +520,7 @@ def create_proto_pop(self):
households, persons, tours = self.expand_template_zones(tables)
households["household_serial_no"] = households[hhid]
else:
households, persons, tours = [self.generate_replicates(k) for k in klist]
households, persons, tours = (self.generate_replicates(k) for k in klist)

# Names
households.name, persons.name, tours.name = klist
Expand Down Expand Up @@ -535,17 +579,17 @@ def inject_tables(self, state: workflow.State):

def annotate_tables(self, state: workflow.State):
# Extract annotations
for annotations in self.model_settings["annotate_proto_tables"]:
tablename = annotations["tablename"]
for annot in self.model_settings.annotate_proto_tables:
tablename = annot.tablename
df = self.state.get_dataframe(tablename)
assert df is not None
assert annotations is not None
assert annot is not None
assign_columns(
state,
df=df,
model_settings={
**annotations["annotate"],
**self.model_settings["suffixes"],
**annot.annotate.dict(),
**self.model_settings.suffixes,
},
trace_label=tracing.extend_trace_label("ProtoPop.annotate", tablename),
)
Expand All @@ -564,7 +608,7 @@ def merge_persons(self):
).merge(
self.land_use,
left_on=self.params["proto_households"]["zone_col"],
right_on=self.model_settings["zone_id_names"]["index_col"],
right_on=self.model_settings.zone_id_names["index_col"],
)

perid = self.params["proto_persons"]["index_col"]
Expand Down Expand Up @@ -597,9 +641,7 @@ def get_disaggregate_logsums(
model_settings = TourLocationComponentSettings.read_settings_file(
state.filesystem, model_name + ".yaml"
)
model_settings.SAMPLE_SIZE = disagg_model_settings.get(
"DESTINATION_SAMPLE_SIZE"
)
model_settings.SAMPLE_SIZE = disagg_model_settings.DESTINATION_SAMPLE_SIZE
estimator = estimation.manager.begin_estimation(state, trace_label)
if estimator:
location_choice.write_estimation_specs(
Expand All @@ -613,7 +655,7 @@ def get_disaggregate_logsums(

# Include the suffix tags to pass onto downstream logsum models (e.g., tour mode choice)
if model_settings.LOGSUM_SETTINGS:
suffixes = util.concat_suffix_dict(disagg_model_settings.get("suffixes"))
suffixes = util.concat_suffix_dict(disagg_model_settings.suffixes)
suffixes.insert(0, str(model_settings.LOGSUM_SETTINGS))
model_settings.LOGSUM_SETTINGS = " ".join(suffixes)

Expand Down
10 changes: 6 additions & 4 deletions activitysim/abm/tables/disaggregate_accessibility.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -161,12 +161,14 @@ def disaggregate_accessibility(state: workflow.State):
disaggregate_accessibility=None,
)

from ..models.disaggregate_accessibility import DisaggregateAccessibilitySettings

# Extract model settings
model_settings = state.filesystem.read_model_settings(
"disaggregate_accessibility.yaml"
model_settings = DisaggregateAccessibilitySettings.read_settings_file(
state.filesystem, "disaggregate_accessibility.yaml"
)
merging_params = model_settings.get("MERGE_ON")
nearest_method = model_settings.get("NEAREST_METHOD", "skims")
merging_params = model_settings.MERGE_ON
nearest_method = model_settings.NEAREST_METHOD
accessibility_cols = [
x for x in proto_accessibility_df.columns if "accessibility" in x
]
Expand Down

0 comments on commit 7bce5f2

Please sign in to comment.