db_from_datafile.py

#!/usr/bin/env python3
"""Load datafile records into TagTracker database.

Options:
    --force: reload given datafiles even if already loaded (default behaviour
            is to successfully load any given file only once, as identified
             by its file fingerprint (md5hash))
    --quiet: no output except error output (to stderr)
    --verbose: extra-chatty output

This should run on the tagtracker server.
Supersedes "tracker2db.py"

This is a script to update a persistent SQLite database in a configurable
directory with data from all available (by default) or specific TagTracker
data files.  It uses some modules that it shares in common with the TT client.

Before running this, the database will need to be created.
See "create_database.sql".

By default this avoids repeating a previously successful datafile load
by keeping a datafile_loads table in the database.  The --force option
will force reloading.

Copyright (C) 2024 Todd Glover & Julias Hocking

    Notwithstanding the licensing information below, this code may not
    be used in a commercial (for-profit, non-profit or government) setting
    without the copyright-holder's written consent.

    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU Affero General Public License as published
    by the Free Software Foundation, either version 3 of the License, or
    (at your option) any later version.

    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU Affero General Public License for more details.

    You should have received a copy of the GNU Affero General Public License
    along with this program.  If not, see <https://www.gnu.org/licenses/>.
"""

import argparse
import glob
import os
import sqlite3
import sys
from dataclasses import dataclass, field
import subprocess
from datetime import datetime
import hashlib

# import tt_datafile
import common.tt_dbutil as db
from common.tt_dbutil import SQL_CRITICAL_ERRORS, SQL_MODERATE_ERRORS

# import tt_globals
import common.tt_util as ut
from common.tt_trackerday import TrackerDay, TrackerDayError
from common.tt_biketag import BikeTagError
from common.tt_bikevisit import BikeVisit

from common.tt_daysummary import DaySummary, PeriodDetail
# from common.tt_time import VTime
from common.tt_constants import REGULAR, OVERSIZE, COMBINED

# Pre-declare this global for linting purposes.
args = None


# Default org_handle -- make this match default text in CREATE_DATABASE sql script
DEFAULT_ORG = "no_org"

# Values for good & bad status
STATUS_GOOD = "GOOD"  # File has not (yet) been rejected or had errors
STATUS_BAD = "BAD"  # File has errors
STATUS_SKIP_GOOD = "SKIP_PRIOR"  # Skip file because was loaded ok previously
STATUS_SKIP_LATER = "SKIP_TAIL"  # Skip file because another same-named is later
STATUS_SKIP_NEWER = "SKIP_NEWER"  # Skip file because another same-named is newer


# Names for tables and columns.s
# Table for logging datafile loads
TABLE_LOADS = "dataloads"
COL_DAYID = "day_id"
COL_FINGERPRINT = "datafile_fingerprint"
COL_DATAFILE = "datafile_name"
COL_TIMESTAMP = "datafile_timestamp"
COL_BATCH = "batch"
## COL_DATE = "date" FIXME: v2 this uses a DAY_ID
# (Also uses COL_DATE and COL_BATCH)

# # Table of individual bike visits
# TABLE_VISITS = "visit"
# COL_ID = "id"  # text date.tag PK
# COL_DATE = "date"
# COL_TAG = "tag"
# COL_BIKETYPE = "type"
# COL_TIME_IN = "time_in"
# COL_TIME_OUT = "time_out"
# COL_DURATION = "duration"
# COL_NOTES = "notes"
# COL_BATCH = "batch"

# # Table of day summaries
# TABLE_DAYS = "day"
# # COL_DATE name reused - text date PK
# COL_REGULAR = "parked_regular"  # int count
# COL_OVERSIZE = "parked_oversize"  # int count
# COL_TOTAL = "parked_total"  # int sum of 2 above
# COL_TOTAL_LEFTOVER = "leftover"  # int count
# COL_MAX_REGULAR = "max_reg"  # int count of max regular bikes
# COL_MAX_REGULAR_TIME = "time_max_reg"  # HHMM time
# COL_MAX_OVERSIZE = "max_over"  # int count of max oversize bikes
# COL_MAX_OVERSIZE_TIME = "time_max_over"  # HHMM time
# COL_MAX_TOTAL = "max_total"  # int sum of 2 above
# COL_MAX_TOTAL_TIME = "time_max_total"  # HHMM
# COL_TIME_OPEN = "time_open"  # HHMM opening time
# COL_TIME_CLOSE = "time_closed"  # HHMM closing time
# COL_DAY_OF_WEEK = "weekday"  # ISO 8601-compliant 1-7 M-S
# COL_PRECIP_MM = "precip_mm"  # mm (bulk pop from EnvCan dat)
# COL_TEMP = "temp"
# COL_SUNSET = "sunset"  # HHMM time at sunset - same
# COL_EVENT = "event"  # brief name of nearby event
# COL_EVENT_PROX = "event_prox_km"  # est. num of km to event
# COL_REGISTRATIONS = "registrations"  # num of bike registrations recorded
# # COL_NOTES name reused
# # COL_BATCH name reused

# # Table of tags contexts
# TABLE_TAGS_CONTEXT = "taglist"
# COL_REGULAR_CONTEXT = "regular"
# COL_OVERSIZE_CONTEXT = "oversize"
# COL_RETIRED_CONTEXT = "retired"

# Bike-type codes. Must match check constraint in code table TYPES.CODE
# REGULAR = "R"
# OVERSIZE = "O"


class DBError(Exception):
    pass


# @dataclass
# class DayStats:
#     # Summary stats & such for one day.
#     # This makes up most of a record for a DAY row,
#     date: str
#     regular_parked: int = 0
#     oversize_parked: int = 0
#     total_parked: int = 0
#     total_leftover: int = 0
#     max_regular_num: int = 0
#     max_regular_time: VTime = ""
#     max_oversize_num: int = 0
#     max_oversize_time: VTime = ""
#     max_total_num: int = 0
#     max_total_time: VTime = ""
#     time_open: VTime = ""
#     time_close: VTime = ""
#     weekday: int = None
#     registrations: int = 0


@dataclass
class Statuses:
    """Keep track of status of individual files & overall program."""

    start_time: str = ut.iso_timestamp()
    status: str = STATUS_GOOD
    errors: int = 0
    error_list: list = field(default_factory=list)
    files: dict = field(default_factory=dict)

    @classmethod
    def set_bad(cls, error_msg: str = "Unspecified error", silent: bool = False):
        cls.status = STATUS_BAD
        cls.errors += 1
        cls.error_list += [error_msg]
        if not silent:
            print(error_msg, file=sys.stderr)

    @classmethod
    def num_files(cls, status: str = "") -> int:
        """Return count of files in a given status.

        status is a known status code, or "".
        If "", returns the total number of files."""
        if status:
            return len([f for f in cls.files.values() if f.status == status])
        return len(cls.files)


@dataclass
class FileInfo:
    name: str
    basename: str = ""
    status: str = STATUS_GOOD
    fingerprint: str = None
    timestamp: str = None
    errors: int = 0
    error_list: list = field(default_factory=list)

    def set_bad(self, error_msg: str = "Unspecified error", silent: bool = False):
        """Set a file's info to a fail state, emit error."""

        def print_first_line(msg: str):
            """Format for printing top line of error list."""
            print(f"{msg} [{self.name}]", file=sys.stderr)

        self.status = STATUS_BAD
        self.errors += 1
        self.error_list += [error_msg]
        if not silent:
            if isinstance(error_msg, list):
                print_first_line(error_msg[0] if error_msg else "")
                for line in error_msg[1:]:
                    print(line, file=sys.stderr)
            else:
                print_first_line(error_msg)


# def create_logtable(dbconx: sqlite3.Connection):
#     """Creates the load-logging table in the database."""
#     if args.verbose:
#         print(f"Assuring table {TABLE_LOADS} exists.")
#     error_msg = sql_exec_and_error(
#         f"""CREATE TABLE IF NOT EXISTS
#         {TABLE_LOADS} (
#             {COL_DATE}  TEXT PRIMARY KEY NOT NULL,
#             {COL_DATAFILE} TEXT NOT NULL,
#             {COL_TIMESTAMP} TEXT NOT NULL,
#             {COL_FINGERPRINT} TEXT NOT NULL,
#             {COL_BATCH} TEXT NOT NULL
#         );""",
#         dbconx,
#     )
#     if error_msg:
#         Statuses.set_bad(f"    SQL error creating datafile_loads: {error_msg}")
#         dbconx.close()
#         sys.exit(1)


def is_linux() -> bool:
    """Check if running in linux."""
    system_platform = sys.platform
    return system_platform.startswith("linux")


def get_file_fingerprint(file_path):
    """Get a file's fingerprint."""

    def get_file_md5_linux(file_path):
        """Get an md5 digest by calling the system program."""
        try:
            result = subprocess.run(
                ["md5sum", file_path], capture_output=True, text=True, check=True
            )
            md5sum_output = result.stdout.strip().split()[0]
            return md5sum_output
        except subprocess.CalledProcessError as e:
            print(f"    Error: {e}", file=sys.stderr)
            return None

    def get_file_md5_windows(file_path):
        """Calculate the MD5 checksum for a given file by reading the file.

        This one would be OS independent.
        """
        # Open the file in binary mode
        with open(file_path, "rb") as file:
            # Create an MD5 hash object
            md5_hash = hashlib.md5()
            # Read the file in chunks to avoid loading the entire file into memory
            for chunk in iter(lambda: file.read(4096), b""):
                # Update the hash object with the chunk
                md5_hash.update(chunk)
        # Get the hexadecimal digest of the hash
        md5_checksum = md5_hash.hexdigest()
        return md5_checksum

    # On linux, prefer calling the system md5sum program
    if is_linux():
        return get_file_md5_linux(file_path)
    else:
        return get_file_md5_windows(file_path)


def get_load_fingerprints(dbconx: sqlite3.Connection) -> list[str]:
    """Get a list of the fingerprints of files last loaded ok."""

    # FIXME: probably want to do for only this organization's scope?

    cursor = dbconx.cursor()
    rows = cursor.execute(f"SELECT {COL_FINGERPRINT} FROM {TABLE_LOADS}").fetchall()
    fingerprints = [r[0] for r in rows]
    cursor.close()

    return fingerprints


def get_file_timestamp(file_path):
    """Get a file's timestamp as an iso8601 string."""
    try:
        timestamp = os.path.getmtime(file_path)
        modified_time = datetime.fromtimestamp(timestamp)
        return modified_time.strftime("%Y-%m-%dT%H:%M:%S")
    except FileNotFoundError:
        print("    File not found {e}.", file=sys.stderr)
        return None


def fetch_existing_weather(
    cursor: sqlite3.Connection.cursor, date: str, orgsite_id: int
) -> tuple[float, float]:
    """Fetch any existing temp and precip data from the given day.

    This uses data and orgsite_id since day_id not known at time this is called."""
    if args.verbose:
        print(f"   Fetching existing weather (temp/precip) for {orgsite_id=}/'{date}'.")
    row = cursor.execute(
        f"SELECT max_temperature, precipitation FROM day WHERE orgsite_id = {orgsite_id} AND date = '{date}';"
    ).fetchone()

    if not row:
        return None, None

    return row[0], row[1]


# FIXME: use versions in tt_dbutil
def fetch_org_id(cursor: sqlite3.Connection.cursor, org_handle: str) -> int:
    """Fetch the org id."""
    org_handle = org_handle.strip().lower()
    if args.verbose:
        print(f"   Fetching org_id for '{org_handle}'.")
    row = cursor.execute(
        f"SELECT id FROM org WHERE org_handle = '{org_handle}';"
    ).fetchone()

    if not row:
        raise DBError(f"No match for org_handle '{org_handle}'.")

    this_id = row[0]
    if not isinstance(this_id, int):
        raise DBError(f"Not integer: org.id '{this_id}' (type {type(this_id)})")
    return this_id


# FIXME: separate making new into its own.  Use fetch_*() from tt_dbutil
def fetch_orgsite_id(
    cursor: sqlite3.Connection.cursor,
    site_handle: str,
    org_id: int,
    site_name: str = "",
    insert_new: bool = False,
) -> int:
    """Fetch the PK id from the orgsite table.

    If not found and insert_new, will create a new record."""

    site_handle = site_handle.strip().lower()
    if args.verbose:
        print(f"   Fetching orgsite_id for '{site_handle}', org_id {org_id}.")

    row = cursor.execute(
        f"SELECT id FROM orgsite WHERE site_handle = '{site_handle}' AND org_id = {org_id};",
    ).fetchone()
    if not row:
        if not insert_new:
            raise DBError(
                f"No match for site_handle '{site_handle}' with org_id {org_id}."
            )
        else:
            this_id = cursor.execute(
                "INSERT INTO orgsite (org_id,site_handle,site_name) VALUES (?,?,?)",
                (org_id, site_handle, site_name),
            ).lastrowid
            return this_id

    this_id = row[0]
    if not isinstance(this_id, int):
        raise DBError(f"Not integer: orgsite.id '{this_id}' (type {type(this_id)})")
    return this_id


def fetch_day_id(
    cursor: sqlite3.Connection.cursor, date: str, orgsite_id: int, null_ok: bool = True
) -> int:
    """Fetch the DAY id for this date/site.

    If null_ok, then it's ok if no matching record; otherwise a DBError."""
    if args.verbose:
        print(f"   Fetching day_id for {orgsite_id=}/'{date}'.")
    row = cursor.execute(
        f"SELECT id FROM day WHERE orgsite_id = {orgsite_id} AND date = '{date}';"
    ).fetchone()

    if not row:
        if null_ok:
            return None
        raise DBError(f"No match for {orgsite_id=}/{date=}.")

    this_id = row[0]
    if not isinstance(this_id, int):
        raise DBError(f"Not integer: day.id '{this_id}' (type {type(this_id)})")

    return this_id


def insert_new_orgsite(
    cursor: sqlite3.Connection.cursor,
    org_id: int,
    site_handle: str,
    site_name: str = "",
) -> int:
    """Insert new org/site into orgsite table, returns orgsite_id."""
    this_id = cursor.execute(
        "INSERT INTO orgsite (org_id,site_handle,site_name) VALUES (?,?,?)",
        (org_id, site_handle, site_name),
    ).lastrowid
    return this_id


def delete_one_day_data(cursor: sqlite3.Connection.cursor, date: str, orgsite_id: int):
    """Delete one complete day's data from all tables."""

    day_id = db.fetch_day_id(cursor=cursor, date=date, maybe_orgsite_id=orgsite_id)

    if args.verbose:
        print(f"   Deleting records for '{orgsite_id=}'/'{date}'.")

    if day_id is not None:
        cursor.execute(f"DELETE FROM VISIT WHERE day_id = {day_id}")
        cursor.execute(f"DELETE FROM BLOCK WHERE day_id = {day_id}")
        cursor.execute(f"DELETE FROM DATALOADS WHERE day_id = {day_id}")
        cursor.execute(f"DELETE FROM DAY WHERE id = {day_id}")


def insert_into_day(
    orgsite_id: int,
    td: TrackerDay,
    summary: DaySummary,
    batch_id: str,
    cursor: sqlite3.Connection.cursor,
) -> int:
    """Load into the DAY table.  Returns the rowid of the record inserted. None if failed."""

    # This requires some fancy footwork in order to preserve any existing
    # environmental values (rain, temp, sunset)
    # since the INSERT OR REPLACE really does do a REPLACE, meaning
    # that (unlike an update), any columns not specifcally names are
    # set to their default values

    # Insert/replace this day's summary info.
    # Have to delete and replace since doing an INSERT OR REPLACE
    # changes the PK id.  So *could* turn constraints off, update the
    # FKs, then turn back on, but this seems simpler since will be adding
    # new VISIT and BLOCK data anyway.  The cost is that must cache the
    # environmental data (temperature and precipitation).

    # Fetch current temp and precip
    existing_temp, existing_precip = fetch_existing_weather(
        cursor=cursor, date=td.date, orgsite_id=orgsite_id
    )

    # Delete existing data for this day
    delete_one_day_data(cursor=cursor, date=td.date, orgsite_id=orgsite_id)

    if args.verbose:
        print("   Adding day summary to database.")

    # Save to DAY table
    cursor.execute(
        """
        INSERT INTO DAY (
            orgsite_id,
            date,
            time_open,
            time_closed,
            weekday,

            num_parked_regular,
            num_parked_oversize,
            num_parked_combined,

            num_remaining_regular,
            num_remaining_oversize,
            num_remaining_combined,

            num_fullest_regular,
            num_fullest_oversize,
            num_fullest_combined,

            time_fullest_regular,
            time_fullest_oversize,
            time_fullest_combined,

            bikes_registered,
            batch,

            max_temperature,
            precipitation
        )
        VALUES (
            ?,?,?,?,?,
            ?,?,?,    ?,?,?,   ?,?,?,   ?,?,?,
            ?,?,
            ?,?
        )
    """,
        (
            # td.org_id,  # Add org_id to TrackerDay class
            # td.site_id,  # Add site_id to TrackerDay class
            orgsite_id,
            summary.whole_day.date,
            summary.whole_day.time_open,
            summary.whole_day.time_closed,
            ut.dow_int(td.date),
            summary.whole_day.num_parked_regular,
            summary.whole_day.num_parked_oversize,
            summary.whole_day.num_parked_combined,
            summary.whole_day.num_remaining_regular,
            summary.whole_day.num_remaining_oversize,
            summary.whole_day.num_remaining_combined,
            summary.whole_day.num_fullest_regular,
            summary.whole_day.num_fullest_oversize,
            summary.whole_day.num_fullest_combined,
            summary.whole_day.time_fullest_regular,
            summary.whole_day.time_fullest_oversize,
            summary.whole_day.time_fullest_combined,
            td.registrations.num_registrations,
            batch_id,
            existing_temp,
            existing_precip,
        ),
    )

    day_id = cursor.lastrowid

    return day_id


# def day_tags_context_into_db(
#     file_info: FileInfo,
#     day: OldTrackerDay,
#     day_totals: DayStats,
#     batch: str,
#     dbconx: sqlite3.Connection,
# ) -> bool:
#     """Load tags context into database."""

#     if args.verbose:
#         print("   Adding lists of retired, regular and oversize tags to database.")

#     regular_tags = ",".join(sorted(list(day.regular)))
#     oversize_tags = ",".join(sorted(list(day.oversize)))
#     retired_tags = ",".join(sorted(list(day.retired)))

#     sql_error = sql_exec_and_error(
#         f"""INSERT OR REPLACE INTO {TABLE_TAGS_CONTEXT} (
#                 {COL_DATE},
#                 {COL_REGULAR_CONTEXT},
#                 {COL_OVERSIZE_CONTEXT},
#                 {COL_RETIRED_CONTEXT}
#             ) VALUES (
#                 '{day_totals.date}',
#                 '{regular_tags}',
#                 '{oversize_tags}',
#                 '{retired_tags}'
#             );""",
#         dbconx,
#     )
#     # Did it work? (No error message means success.)
#     if sql_error:
#         file_info.set_bad(f"SQL error adding to {TABLE_TAGS_CONTEXT}: {sql_error}")
#         return False

#     return True


def insert_into_visit(
    day: TrackerDay,
    cursor: sqlite3.Connection.cursor,
    day_id: int,
) -> bool:
    """Load this day's visits into the database."""
    # if args.verbose:
    #     print("   Deleting any existing visits info for this day from database.")
    # sql_error = sql_exec_and_error(
    #     f"DELETE FROM {TABLE_VISITS} WHERE date = '{day_totals.date}'", dbconx
    # )
    # if sql_error:
    #     file_info.set_bad(f"Error deleting VISIT rows: {sql_error}")
    #     return False

    if args.verbose:
        print(f"   Adding {len(day.all_visits())} visits to database.")

    for visit in day.all_visits():
        visit: BikeVisit
        biketype = "R" if day.biketags[visit.tagid].bike_type == REGULAR else "O"

        # Save to VISIT table
        cursor.execute(
            """
            INSERT INTO VISIT (
                day_id,
                time_in,
                time_out,
                duration,
                bike_type,
                bike_id
            )

            VALUES (?,?,?,?,?,?)
        """,
            (
                day_id,
                visit.time_in,
                visit.time_out,
                visit.duration(day.time_closed,is_close_of_business=True),
                biketype,
                visit.tagid,
            ),
        )
    return True


def insert_into_block(
    summary: DaySummary,
    cursor: sqlite3.Connection.cursor,
    day_id: int,
) -> bool:
    """Load this day's block data into the database."""

    if args.verbose:
        print(f"   Adding {len(summary.blocks)} data blocks to database.")

    for block in summary.blocks.values():
        block: PeriodDetail
        # Save to BLOCK table
        cursor.execute(
            """
            INSERT INTO block (

                day_id,
                time_start,

                num_incoming_regular,
                num_incoming_oversize,
                num_incoming_combined,

                num_outgoing_regular,
                num_outgoing_oversize,
                num_outgoing_combined,

                num_on_hand_regular,
                num_on_hand_oversize,
                num_on_hand_combined,

                num_fullest_regular,
                num_fullest_oversize,
                num_fullest_combined,

                time_fullest_regular,
                time_fullest_oversize,
                time_fullest_combined


            )

            VALUES (?,?,  ?,?,?,  ?,?,?,  ?,?,?,  ?,?,?, ?,?,?)
        """,
            (
                day_id,
                block.time_start,
                block.num_incoming[REGULAR],
                block.num_incoming[OVERSIZE],
                block.num_incoming[COMBINED],
                block.num_outgoing[REGULAR],
                block.num_outgoing[OVERSIZE],
                block.num_outgoing[COMBINED],
                block.num_on_hand[REGULAR],
                block.num_on_hand[OVERSIZE],
                block.num_on_hand[COMBINED],
                block.num_fullest[REGULAR],
                block.num_fullest[OVERSIZE],
                block.num_fullest[COMBINED],
                block.time_fullest[REGULAR],
                block.time_fullest[OVERSIZE],
                block.time_fullest[COMBINED],
            ),
        )
    return True


def insert_into_dataloads(
    file_info: FileInfo, day_id: int, batch: str, dbconx: sqlite3.Connection
) -> bool:
    """Update table that tracks load successes."""
    if args.verbose:
        print(f"   Updating fileload info for {file_info.basename} into database.")
    sql = f"""
        INSERT OR REPLACE INTO {TABLE_LOADS} (
            {COL_DAYID},
            {COL_DATAFILE},
            {COL_TIMESTAMP},
            {COL_FINGERPRINT},
            {COL_BATCH}
        ) VALUES (
            '{day_id}',
            '{file_info.name}',
            '{file_info.timestamp}',
            '{file_info.fingerprint}',
            '{batch}'
        ) """
    cursor = dbconx.cursor()
    try:
        cursor.execute(sql)
    except tuple(SQL_MODERATE_ERRORS) as e:
        print(f"Non-fatal error occurred: {e}")
        return False
    except tuple(SQL_CRITICAL_ERRORS) as e:
        print(f"Fatal error occurred: {e}")
        raise
    finally:
        cursor.close()

    return True


def read_datafile(filename: str) -> tuple[TrackerDay, DaySummary]:
    """Read and prepare one filename into TrackerDay & DaySummary obects."""
    file_info: FileInfo = Statuses.files[filename]

    if not os.path.isfile(filename):
        file_info.set_bad("File not found")
        return None, None

    if args.verbose:
        print(f"Reading {filename}:")

    try:
        day = TrackerDay.load_from_file(filename)
    except (TrackerDayError, BikeTagError) as e:
        msg = f"Error reading file {filename=}: {e}. Skipping file."
        file_info.set_bad(msg)
        # print(msg)
        return None, None

    lint_msgs = day.lint_check(strict_datetimes=True, allow_quick_checkout=True)
    if lint_msgs:
        msg = [f"Errors reading datafile {filename}:"] + lint_msgs
        file_info.set_bad(msg)
        return None, None

    day_summary = DaySummary(day=day, as_of_when="24:00")
    return day, day_summary


def one_datafile_into_db(filename: str, batch, dbconx) -> str:
    """Record one datafile to the database, returns date.

    Reads the datafile, looking for errors.
    Calculates summary stats, tables.
    Saves the file info & fingerprint to LOAD_INFO table.

    Returns the date of the file's data if successful (else "")
    Status is also Statuses.files[filename], which is a FileInfo object.

    This should create then destroy a cursor, which means rollback or commit.
    """

    day, day_summary = read_datafile(filename=filename)
    if day is None or day_summary is None:
        return None

    # Datafile fully loaded, now start db stuff
    cursor = sql_begin_transaction(dbconx=dbconx)
    if args.verbose:
        print("   Fetching org and orgsize ids")
    org_id = db.fetch_org_id(cursor=cursor, org_handle=args.org)
    if org_id is None:
        raise DBError(f"No org matches {args.org_handle}")
    orgsite_id = db.fetch_orgsite_id(
        cursor=cursor, site_handle=day.site_handle, maybe_org_id=org_id, null_ok=True
    )
    if orgsite_id is None:
        if args.verbose:
            print(f"  Adding new orgsite {args.org=},{day.site_handle=}")
        orgsite_id = insert_new_orgsite(
            cursor=cursor,
            org_id=org_id,
            site_handle=day.site_handle,
            site_name=day.site_name,
        )

    if args.verbose:
        print(
            f"   Date:{day.date}  Open:{day.time_open}-{day.time_closed}"
            f"  Visits:{day_summary.whole_day.num_parked_combined}  "
            f"Leftover:{day_summary.whole_day.num_remaining_combined}  "
            f"Registrations:{day.registrations.num_registrations}"
        )

    # Save data to database
    try:

        day_id = insert_into_day(
            td=day,
            summary=day_summary,
            orgsite_id=orgsite_id,
            batch_id=batch,
            cursor=cursor,
        )
        insert_into_visit(
            day=day,
            cursor=cursor,
            day_id=day_id,
        )
        insert_into_block(
            summary=day_summary,
            cursor=cursor,
            day_id=day_id,
        )
        # day_tags_context_into_db(file_info, day, day_totals, batch, dbconx)
        # day_blocks_into_db()

        insert_into_dataloads(
            Statuses.files[filename], day_id=day_id, batch=batch, dbconx=dbconx
        )

    except (sqlite3.Error, DBError) as err:
        print(f"Error:{err}", file=sys.stderr)
        dbconx.rollback()
        return ""

    finally:
        cursor.close()

    dbconx.commit()  # commit as one datafile transaction

    if args.verbose:
        print(
            f"   Committed {day_summary.whole_day.num_parked_combined} visits for {day.date}."
        )

    return day.date


def sql_begin_transaction(dbconx: sqlite3.Connection) -> sqlite3.Connection.cursor:
    curs = dbconx.cursor()
    curs.execute("BEGIN;")
    return curs


def sql_exec_and_error(sql_statement: str, dbconx) -> int:
    """Execute a SQL statement, returns error message (if any).

    Presumably there is no error if the return is empty.
    Counter-intuitively this means that this returns bool(False) if
    this succeeded.

    Not suitable (of course) for SELECT statements.
    """
    try:
        curs = dbconx.cursor()
        curs.execute(sql_statement)
    except sqlite3.Error as sqlite_err:
        print(f"SQLite error:{sqlite_err}", file=sys.stderr)
        return sqlite_err if sqlite_err else "Unspecified SQLite error"
    # Success
    return ""


def get_args() -> argparse.Namespace:
    """Get program arguments."""

    parser = argparse.ArgumentParser()

    parser.add_argument(
        "-q", "--quiet", action="store_true", help="suppresses all non-error output"
    )
    parser.add_argument(
        "-v", "--verbose", action="store_true", help="provides most detailed output"
    )
    parser.add_argument(
        "-f", "--force", action="store_true", help="load even if previously loaded"
    )
    parser.add_argument(
        "-n",
        "--newest-only",
        action="store_true",
        help="if files have same basename, only consider the newest one",
    )
    parser.add_argument(
        "-t",
        "--tail-only",
        action="store_true",
        help="if files have same basename, only consider "
        "the one that is closer to the tail of of the file list",
    )
    parser.add_argument(
        "-o",
        "--org",
        type=str,
        default=DEFAULT_ORG,
        help=f"Organization org_handle (default: '{DEFAULT_ORG}')",
    )
    parser.add_argument(
        "dataglob", type=str, nargs="+", help="Fileglob(s) of datafiles to load"
    )
    parser.add_argument("dbfile", type=str, help="SQLite3 database file")

    prog_args = parser.parse_args()

    if prog_args.verbose and prog_args.quiet:
        prog_args.quiet = False
        print("Arg --verbose set; ignoring arg --quiet.")

    if prog_args.tail_only and prog_args.newest_only:
        print(
            "Error: Can not use --newest-only and --tail-only together", file=sys.stderr
        )
        sys.exit(1)

    return prog_args


def find_datafiles(fileglob: list) -> list:
    """Return a list of files from a dataglob." """

    maybe_datafiles = []
    for this_glob in fileglob:
        # print(f"{this_glob=}")
        globbed_files = glob.glob(this_glob)
        if len(globbed_files) < 1:
            print(
                f"Error: no files found matching glob '{this_glob}'",
                file=sys.stderr,
            )
            sys.exit(1)
        maybe_datafiles += globbed_files
    # Make paths absolute
    maybe_datafiles = convert_paths_to_absolute(maybe_datafiles)
    # Exclude exact duplicates but preserve the file order
    maybe_datafiles = dedup_filepaths(maybe_datafiles)

    return maybe_datafiles


def dedup_filepaths(filepaths: list[str]) -> list[str]:
    """Deduplicate a list of filepaths, removing dups and preserving order."""
    dedupped = []
    dups = []
    for f in filepaths:
        if f in dedupped:
            if f not in dups and not args.quiet:
                print(
                    f"Warning: filepath {f} given more than once, ignoring later instances.",
                    file=sys.stderr,
                )
            dups.append(f)
        else:
            dedupped.append(f)
    return dedupped


def convert_paths_to_absolute(files: list) -> list:
    """Convert a list of relative paths to absolute paths."""
    abs_paths = []
    for f in files:
        abs_paths.append(os.path.abspath(f))
    return abs_paths


def get_files_metadata(maybe_datafiles: list):
    """Gets metadata for the files, loads it into Statuses class."""
    for f in maybe_datafiles:
        f_info = FileInfo(f)
        Statuses.files[f] = f_info
        if not os.path.exists(f):
            f_info.set_bad("File not found")
            continue
        f_info.basename = os.path.basename(f)
        f_info.fingerprint = get_file_fingerprint(f)
        if not f_info.fingerprint:
            f_info.set_bad("Can not read md5sum")
            continue
        f_info.timestamp = get_file_timestamp(f)
        if not f_info.timestamp:
            f_info.set_bad("Can not read timestamp")
            continue

    ok_datafiles = [
        f for f in Statuses.files if Statuses.files[f].status == STATUS_GOOD
    ]

    return ok_datafiles


def skip_non_tail_dups(filepath_list):
    """Sets FileInfo for some files based on their location in the list.

    file_list is an ordered list of filepaths.
    """
    base_list = [os.path.basename(f) for f in filepath_list]
    for i, base in enumerate(
        base_list[:-1]
    ):  # Iterate over all elements except the last one
        if (
            base in base_list[i + 1 :]
        ):  # Check if the item appears again in the subsequent portion of the list
            skip_path = filepath_list[i]
            # Set this file's status to skip
            Statuses.files[skip_path].status = STATUS_SKIP_LATER
    if args.verbose:
        print(
            f"{Statuses.num_files(STATUS_SKIP_LATER)} same-named files ignored "
            "because another is later in arg list."
        )


def skip_non_newest_dups():
    """Skips duplicate files that are not the newest."""

    def get_fileinfo_timestamp(fi: FileInfo) -> str:
        """FileInfo timestamp. (as a function so can use as key to max())."""
        return fi.timestamp

    # Find all the fileinfos associated with each basename
    bases = {}
    for fi in Statuses.files.values():
        fi: FileInfo
        base = fi.basename
        if base not in bases:
            bases[base] = [fi]
        else:
            bases[base].append(fi)
    for fi_list in bases.values():
        if len(fi_list) > 1:
            # Find newest file, skip the rest
            newest_stamp = max(fi_list, key=get_fileinfo_timestamp)
            for dup_fi in fi_list:
                dup_fi: FileInfo
                if dup_fi.timestamp != newest_stamp.timestamp:
                    dup_fi.status = STATUS_SKIP_NEWER
    if args.verbose:
        print(
            f"{Statuses.num_files(STATUS_SKIP_NEWER)} same-named files ignored because another is newer."
        )


def print_summary():
    """A chatty summary of what took place."""
    for info in Statuses.files.values():
        info: FileInfo
        info.errors = info.errors if info.errors else len(info.error_list)

    print()
    print(f"{Statuses.num_files():4d} files requested")
    print(f"{Statuses.num_files(STATUS_GOOD):4d} files loaded OK")
    print(
        f"{Statuses.num_files(STATUS_SKIP_LATER):4d} files ignored, a file with same name is later in file list"
    )
    print(
        f"{Statuses.num_files(STATUS_SKIP_NEWER):4d} files ignored, a file with same name is newer"
    )
    print(
        f"{Statuses.num_files(STATUS_SKIP_GOOD):4d} files ignored, previously loaded ok"
    )
    print(f"{Statuses.num_files(STATUS_BAD):4d} files not loaded because of errors")
    if args.verbose:
        for f, finfo in sorted(Statuses.files.items()):
            finfo: FileInfo
            if finfo.status == STATUS_BAD:
                print(f"    {f}: {finfo.errors} errors")


def main():
    """Main routine."""

    Statuses.files = {}

    global args  # pylint:disable=global-statement
    args = get_args()
    ordered_file_list = find_datafiles(args.dataglob)
    if not args.quiet:
        print(f"Load requested for {len(ordered_file_list)} files.")

    if args.verbose:
        print("Calculating metadata of datafiles.")
    get_files_metadata(ordered_file_list)

    database_file = args.dbfile
    if args.verbose:
        print(f"Connecting to database {database_file}.")
    dbconx = db.db_connect(database_file)
    if not dbconx:
        print(f"Error: unable to connect to db {database_file}", file=sys.stderr)
        try:
            dbconx.close()
        except Exception as e:
            print(f"Error {e}")
        sys.exit(1)

    if args.tail_only:
        skip_non_tail_dups(ordered_file_list)

    if args.newest_only:
        skip_non_newest_dups()

    # Get a list of fingerprints of previous good loads.
    if args.verbose:
        print("Fetching fingerprints of previous datafile loads.")
    try:
        good_fingerprints = get_load_fingerprints(dbconx=dbconx)
    except tuple(SQL_MODERATE_ERRORS + SQL_CRITICAL_ERRORS) as e:
        print(f"SQL error: {e}", file=sys.stderr)
        sys.exit(1)

    # Decide which files to ignore becuase previously loaded ok
    for finfo in Statuses.files.values():
        finfo: FileInfo
        if (
            not args.force
            and finfo.status == STATUS_GOOD
            and finfo.fingerprint in good_fingerprints
        ):
            finfo.status = STATUS_SKIP_GOOD

    if not args.quiet:
        num_skipped = Statuses.num_files(STATUS_SKIP_GOOD)
        # ut.squawk(f"{num_skipped=},{args.force=},{args.quiet=}")
        if num_skipped:
            if args.force:
                print(f"Forcing reload of {num_skipped} previously loaded files.")
            else:
                print(f"Ignoring {num_skipped} previously loaded files.")
        num_bad = Statuses.num_files(STATUS_BAD)
        if num_bad:
            print(f"Ignoring {num_bad} files already known to be bad.")
        if num_skipped or num_bad:
            print(
                f"Attempting to load {Statuses.num_files(STATUS_GOOD)} remaining files."
            )

    num_good = Statuses.num_files(STATUS_GOOD)

    if num_good:
        if args.verbose:
            print("Assuring foreign key constraints are enabled.")
        try:
            cursor = dbconx.cursor()
            cursor.execute("PRAGMA foreign_keys=ON;")
            dbconx.commit()
        except tuple(SQL_MODERATE_ERRORS) as e:
            print(f"SQL error: {e}", file=sys.stderr)
            sys.exit(1)
        except tuple(SQL_CRITICAL_ERRORS) as e:
            print(f"Serious SQL error occurred: {e}", file=sys.stderr)
            raise
        finally:
            cursor.close()

        batch = Statuses.start_time[
            :-3
        ]  # For some reason batch does not include seconds
        if args.verbose:
            print(f"Batch ID is {batch}.")

    # Load the datafiles.
    for file_name in sorted(
        [f.name for f in Statuses.files.values() if f.status == STATUS_GOOD]
    ):
        try:
            one_datafile_into_db(file_name, batch, dbconx)

        except tuple(SQL_MODERATE_ERRORS) as e:
            print(f"SQL error: {e}")
            continue
        except tuple(SQL_CRITICAL_ERRORS) as e:
            print(f"Serious SQL error occurred: {e}")
            raise

    if args.verbose:
        print("Closing database connection.")
    dbconx.close()

    if not args.quiet:
        print_summary()

    if Statuses.errors:
        sys.exit(1)
    sys.exit(0)


if __name__ == "__main__":
    main()