mp

gistfile1.txt

* `docker container run --rm -d --name mdb mongo`
* `docker container inspect mdb | grep IPA` # to get IP for code unless this is the only container
* `python3 -m venv faster`
* cd `faster`
* `source ./bin/activate`
* `python3 -m pip install pymongo

To run the code:

* `python3 code.py #writers #readers #docs`
* `python3 code.py 10 10 1000`

The prior command will run 10 inserting processes adding 1000 documents a piece; then 10 querying processes reading the 1000 docs just inserted.

Below is the code.py.

```
import multiprocessing as mp
from pymongo import MongoClient
import datetime
import sys

def drop():
    client = MongoClient('172.17.0.2', 27017)
    db = client.test_database
    collection = db.test_collection
    while collection.drop():
        print("still here")
        pass

    print("drop done")

def insert(t):
    x, y = t
    print(f"{x} started inserting")
    client = MongoClient('172.17.0.2', 27017)
    db = client.test_database
    collection = db.test_collection
    dedup = chr(x + 100)

    for i in range(y):
        id = str(x) + str(i) + dedup 
        post = {"author": "Mike",
        "text": "My first blog post!",
        "tags": ["mongodb", "python", "pymongo"],
        "date": datetime.datetime.utcnow(),
        "_id": id}
        collection.insert_one(post)
        if i % 100:
            print(f"inserted {x}{id}")

    print(f"{x} finished inserting")

def query(t):
    x, y = t
    print(f"{x} started querying")
    client = MongoClient('172.17.0.2', 27017)
    db = client.test_database
    collection = db.test_collection
    dedup = chr(x + 100)

    for i in range(y):
        id = str(x) + str(i) + dedup 
        result = collection.find_one({"_id": id})
        if i % 100:
            print(f"retrieved {x}{id}:{result}")

    print(f"{x} finished query")

if __name__ == '__main__':
    num_insert_procs = int(sys.argv[1])
    num_read_procs = int(sys.argv[2])
    doc_count = int(sys.argv[3])

    drop()

    with mp.get_context('spawn').Pool() as p1:
        l = list(range(1, num_insert_procs+1))
        p1.map(insert, list(((i, doc_count) for i in l)))

    with mp.get_context('spawn').Pool() as p2:
        l = list(range(1, num_read_procs+1))
        p2.map(query, list(((i, doc_count) for i in l)))
```