ampo 0.3.0rc2

Python library async ORM for mongodb. Object as pyDantic

Aio Mongo Pydantic ORM (ampo)

Features:

• Asynchronous
• Many to many relationships
• Support MongoDB from 4.2
• Python 3.8+

Usage

All example run into:

python -m asyncio

Create and get object

from ampo import CollectionWorker, AMPODatabase, ORMConfig, init_collection

# Initilize DB before calls db methods
AMPODatabase(url="mongodb://test")

# Pydantic Model
class ModelA(CollectionWorker):
    field1: str
    field2: int

    model_config = ORMConfig(
        orm_collection="test"
    )

await init_collection()

inst_a = ModelA("test", 123)
await inst_a.save()

# Get object
inst_a = await ModelA.get(field1="test")

Get all objects

Support additional options for this method. See find().

Example:

from ampo import CollectionWorker, AMPODatabase, ORMConfig, init_collection

# Initilize DB before calls db methods
AMPODatabase(url="mongodb://test")

# Pydantic Model
class ModelA(CollectionWorker):
    field1: str
    field2: int
    
    model_config = ORMConfig(
        orm_collection="test"
    )

await init_collection()

inst_a = ModelA("test", 123)
await inst_a.save()

# Get all objects
insts = await ModelA.get_all()

# Get all objects, by filter, and addional options
insts = await ModelA.get_all(
    filter={"field1": "test"},
    sort=[("field2", 1)],
    limit=10,
    skip=0
)

Id

For search by 'id' usages in filter '_id' or 'id' name.

Indexes

# import

# Initilize DB before calls db methods
AMPODatabase(url="mongodb://test")

# Pydantic Model
class ModelA(CollectionWorker):
    field1: str

    model_config = ORMConfig(
        orm_collection="test",
        orm_indexes=[
            {
                "keys": ["field1"],
                "options": {
                    "unique": True
                }
            }
        ]
    )

# This method create indexes
# Call only one time
await init_collection()

Suppport options:

• unique
• expireAfterSeconds

Keys is list of fields.

TTL Index

It works only with single field (TTL Indexes).

You should set the option 'expireAfterSeconds', and field 'keys' should have only single field.

Example:

# import

# Initilize DB before calls db methods
AMPODatabase(url="mongodb://test")

# Pydantic Model
class ModelA(CollectionWorker):
    field1: datetime

    model_config = ORMConfig(
        orm_collection="test",
        orm_indexes=[
            {
                "keys": ["field1"],
                "options": {
                    "expireAfterSeconds": 10
                }
            }
        ]
    )

# optional, set new value
ModelA.update_expiration_value("field1", 20)

await init_collection()

if you want to set the 'expireAfterSeconds' only from method 'update_expiration_value', set it to '-1'. if you want skip the index changed, call method 'expiration_index_skip' before init_collection.

Indexes in replica set cluster

The replica set cluster has a specific behavior when creating indexes. If one of the nodes in the cluster is not reachable, the index creation will wait for the node to become available. See Index Builds in Replicated Environments. Change this behavior by setting the 'commit_quorum' option to 'majority'. See createIndexes.

Supported only from MongoDB version 4.4.

Example:

# import

# Initilize DB before calls db methods
AMPODatabase(url="mongodb://test")

# Pydantic Model
class ModelA(CollectionWorker):
    field1: str

    model_config = ORMConfig(
        orm_collection="test",
        orm_indexes=[
            {
                "keys": ["field1"],
                "commit_quorum": "majority",
            }
        ]
    )

await init_collection()

Relationships between documents

Embeded

It is supported by default. Just, you need create the embedded document as class of pydantic - 'BaseModel'. It will be stored into db as object.

Example:

from pydantic import BaseModel

class Embeded(BaseModel):
    name: str

class ModelA(CollectionWorker):
    field1: str
    field2: Embeded

    model_config = ORMConfig(
        orm_collection="test"
    )

Lock Record

It is a mechanism that allows you to retrieve a record with a lock. It is based on the findOneAndUpdate(). When the record is found, the field "lock_field" is set to True. And when the next search is made, this record will be skipped.

Example:

import datetime
from typing import Optional
from ampo import CollectionWorker, AMPODatabase, ORMConfig, init_collection

# Pydantic Model
class ModelA(CollectionWorker):
    field1: str
    lfield: bool = False
    field_dt_start: Optional[datetime.datetime] = None

    model_config = ORMConfig(
        orm_collection="test",
        orm_lock_record={
            "lock_field": "lfield",
            "lock_field_time_start": "field_dt_start",
        }
    )

await init_collection()

inst_a = ModelA("test", 123)
await inst_a.save()

inst_a = await ModelA.get_and_lock(field1="test")
# process
await inst_a.reset_lock()

# as context
async with ModelA.get_and_lock_context(field1="test") as inst_a:
    pass
    # process

# as context, version II
# timeout on wait
try:
  async with ModelA.get_lock_wait_context(
      filter={"field1": "test"}, timeout=10
  ) as inst_a:
      pass
except (TimeoutError, ValueError) as e:
    print("Error:", e)

Mehanism reset lock. If lock exist more than time, set 'lock_max_period_sec', lock will be reset. Default value is 15 minutes.

Development

Style:

• NumPy/SciPy docstrings style guide

Run tests:

env TEST_MONGO_URL=mongodb://localhost/test pytest