path: root/qc_app/samples.py

"""Code regarding samples"""
import os
import sys
import csv
import uuid
from pathlib import Path
from typing import Iterator

import MySQLdb as mdb
from redis import Redis
from MySQLdb.cursors import DictCursor
from flask import (
    flash,
    request,
    url_for,
    redirect,
    Blueprint,
    render_template,
    current_app as app)

from quality_control.parsing import take

from qc_app import jobs
from qc_app.files import save_file
from qc_app.dbinsert import species_by_id, groups_by_species
from qc_app.db_utils import (
    with_db_connection,
    database_connection,
    with_redis_connection)

samples = Blueprint("samples", __name__)

@samples.route("/upload/species", methods=["POST"])
def select_species():
    """Select the species."""
    index_page = redirect(url_for("entry.upload_file"))
    species_id = request.form.get("species_id")
    if bool(species_id):
        species_id = int(species_id)
        species = with_db_connection(
            lambda conn: species_by_id(conn, species_id))
        if bool(species):
            return render_template(
                "samples/select-population.html",
                species=species,
                populations=groups_by_species(species_id))
        flash("Invalid species selected!", "alert-error")
    flash("You need to select a species", "alert-error")
    return index_page

@samples.route("/upload/create-population", methods=["POST"])
def create_population():
    """Create new grouping/population."""
    species_page = redirect(url_for("samples.select_species"), code=307)
    with database_connection(app.config["SQL_URI"]) as conn:
        species = species_by_id(conn, request.form.get("species_id"))
        pop_name = request.form.get("inbredset_name").strip()
        pop_fullname = request.form.get("inbredset_fullname").strip()

        if not bool(species):
            flash("Invalid species!", "alert-error error-create-population")
            return species_page
        if (not bool(pop_name)) or (not bool(pop_fullname)):
            flash("You *MUST* provide a grouping/population name",
                  "alert-error error-create-population")
            return species_page

        pop_id = save_population(conn, {
            "SpeciesId": species["SpeciesId"],
            "Name": pop_name,
            "InbredSetName": pop_fullname,
            "FullName": pop_fullname,
            "Family": request.form.get("inbredset_family") or None,
            "Description": request.form.get("description") or None
        })

    flash("Grouping/Population created successfully.", "alert-success")
    return render_template(
        "samples/upload-samples.html",
        species=species,
        population=with_db_connection(
            lambda conn: population_by_id(conn, pop_id)))

@samples.route("/upload/select-population", methods=["POST"])
def select_population():
    """Select from existing groupings/populations."""
    species_page = redirect(url_for("samples.select_species"), code=307)
    with database_connection(app.config["SQL_URI"]) as conn:
        species = species_by_id(conn, request.form.get("species_id"))
        pop_id = int(request.form.get("inbredset_id"))
        population = with_db_connection(lambda conn: population_by_id(conn, pop_id))

    if not bool(species):
        flash("Invalid species!", "alert-error error-select-population")
        return species_page

    if not bool(population):
        flash("Invalid grouping/population!",
              "alert-error error-select-population")
        return species_page

    return render_template("samples/upload-samples.html",
                           species=species,
                           population=population)

def read_samples_file(filepath, separator: str, firstlineheading: bool, **kwargs) -> Iterator[dict]:
    """Read the samples file."""
    with open(filepath, "r", encoding="utf-8") as inputfile:
        reader = csv.DictReader(
            inputfile,
            fieldnames=(
                None if firstlineheading
                else ("Name", "Name2", "Symbol", "Alias")),
            delimiter=separator,
            quotechar=kwargs.get("quotechar", '"'))
        for row in reader:
            yield row

def save_samples_data(conn: mdb.Connection,
                      speciesid: int,
                      file_data: Iterator[dict]):
    """Save the samples to DB."""
    data = ({**row, "SpeciesId": speciesid} for row in file_data)
    total = 0
    with conn.cursor() as cursor:
        while True:
            batch = take(data, 5000)
            if len(batch) == 0:
                break
            cursor.executemany(
                "INSERT INTO Strain(Name, Name2, SpeciesId, Symbol, Alias) "
                "VALUES("
                "    %(Name)s, %(Name2)s, %(SpeciesId)s, %(Symbol)s, %(Alias)s"
                ") ON DUPLICATE KEY UPDATE Name=Name",
                batch)
            total += len(batch)
            print(f"\tSaved {total} samples total so far.")

def cross_reference_samples(conn: mdb.Connection,
                            species_id: int,
                            population_id: int,
                            strain_names: Iterator[str]):
    """Link samples to their population."""
    with conn.cursor(cursorclass=DictCursor) as cursor:
        cursor.execute(
            "SELECT MAX(OrderId) AS loid FROM StrainXRef WHERE InbredSetId=%s",
            (population_id,))
        last_order_id = (cursor.fetchone()["loid"] or 10)
        total = 0
        while True:
            batch = take(strain_names, 5000)
            if len(batch) == 0:
                break
            params_str = ", ".join(["%s"] * len(batch))
            ## This query is slow -- investigate.
            cursor.execute(
                "SELECT s.Id FROM Strain AS s LEFT JOIN StrainXRef AS sx "
                "ON s.Id = sx.StrainId WHERE s.SpeciesId=%s AND s.Name IN "
                f"({params_str}) AND sx.StrainId IS NULL",
                (species_id,) + tuple(batch))
            strain_ids = (sid["Id"] for sid in cursor.fetchall())
            params = tuple({
                "pop_id": population_id,
                "strain_id": strain_id,
                "order_id": last_order_id + (order_id * 10),
                "mapping": "N",
                "pedigree": None
            } for order_id, strain_id in enumerate(strain_ids, start=1))
            cursor.executemany(
                "INSERT INTO StrainXRef( "
                "  InbredSetId, StrainId, OrderId, Used_for_mapping, PedigreeStatus"
                ")"
                "VALUES ("
                "  %(pop_id)s, %(strain_id)s, %(order_id)s, %(mapping)s, "
                "  %(pedigree)s"
                ")",
                params)
            last_order_id += (len(params) * 10)
            total += len(batch)
            print(f"\t{total} total samples cross-referenced to the population "
                  "so far.")

def build_sample_upload_job(# pylint: disable=[too-many-arguments]
        speciesid: int,
        populationid: int,
        samplesfile: Path,
        separator: str,
        firstlineheading: bool,
        quotechar: str):
    """Define the async command to run the actual samples data upload."""
    return [
        sys.executable, "-m", "scripts.insert_samples", app.config["SQL_URI"],
        str(speciesid), str(populationid), str(samplesfile.absolute()),
        separator, f"--redisuri={app.config['REDIS_URL']}",
        f"--quotechar={quotechar}"
    ] + (["--firstlineheading"] if firstlineheading else [])

@samples.route("/upload/samples", methods=["POST"])
def upload_samples():
    """Upload the samples."""
    samples_uploads_page = redirect(url_for("samples.select_population"),
                                    code=307)

    with database_connection(app.config["SQL_URI"]) as conn:
        species = species_by_id(conn, request.form.get("species_id"))
        if not bool(species):
            flash("Invalid species!", "alert-error")
            return samples_uploads_page

        population = with_db_connection(
            lambda conn: population_by_id(
                conn, int(request.form.get("inbredset_id"))))
        if not bool(population):
            flash("Invalid grouping/population!", "alert-error")
            return samples_uploads_page

        samples_file = save_file(request.files["samples_file"], Path(app.config["UPLOAD_FOLDER"]))
        if not bool(samples_file):
            flash("You need to provide a file with the samples data.")
            return samples_uploads_page

        firstlineheading = (request.form.get("first_line_heading") == "on")

        separator = request.form.get("separator")
        if separator == "other":
            separator = request.form.get("other_separator")
        if not bool(separator):
            flash("You need to provide a separator character.", "alert-error")
            return samples_uploads_page

        quotechar = (request.form.get("field_delimiter", '"') or '"')

        redisuri = app.config["REDIS_URL"]
        with Redis.from_url(redisuri, decode_responses=True) as rconn:
            the_job = jobs.launch_job(
                jobs.initialise_job(
                    rconn,
                    str(uuid.uuid4()),
                    build_sample_upload_job(
                        species["SpeciesId"],
                        population["InbredSetId"],
                        samples_file,
                        separator,
                        firstlineheading,
                        quotechar),
                    "samples_upload",
                    app.config["JOBS_TTL_SECONDS"],
                    {"job_name": f"Samples Upload: {samples_file.name}"}),
                redisuri,
                f"{app.config['UPLOAD_FOLDER']}/job_errors")
            return redirect(url_for(
                "samples.upload_status", job_id=the_job["job_id"]))

@samples.route("/upload/status/<uuid:job_id>", methods=["GET"])
def upload_status(job_id: uuid.UUID):
    """Check on the status of a samples upload job."""
    job = with_redis_connection(lambda rconn: jobs.job(rconn, str(job_id)))
    if job:
        status = job["status"]
        if status == "success":
            return render_template("samples/upload-success.html", job=job)

        if status == "error":
            return redirect(url_for("samples.upload_failure", job_id=job_id))

        error_filename = Path(jobs.error_filename(
            job_id, f"{app.config['UPLOAD_FOLDER']}/job_errors"))
        if error_filename.exists():
            stat = os.stat(error_filename)
            if stat.st_size > 0:
                return redirect(url_for(
                    "samples.upload_failure", job_id=job_id))

        return render_template(
            "samples/upload-progress.html",
            job=job) # maybe also handle this?

    return render_template("no_such_job.html", job_id=job_id), 400

@samples.route("/upload/failure/<uuid:job_id>", methods=["GET"])
def upload_failure(job_id: uuid.UUID):
    """Display the errors of the samples upload failure."""
    job = with_redis_connection(lambda rconn: jobs.job(rconn, str(job_id)))
    if not bool(job):
        return render_template("no_such_job.html", job_id=job_id), 400

    error_filename = Path(jobs.error_filename(
        job_id, f"{app.config['UPLOAD_FOLDER']}/job_errors"))
    if error_filename.exists():
        stat = os.stat(error_filename)
        if stat.st_size > 0:
            return render_template("worker_failure.html", job_id=job_id)

    return render_template("samples/upload-failure.html", job=job)