anonlink-entity-service

anonlink-entity-service copied to clipboard

I experienced some issues on a first time install on a Windows machine using v.1.15.1. A summary of what I did to resolve these issues is shown below as are...

greshje

Replace low level database operations with an ORM

2

I quite like the look of [peewee](http://peewee.readthedocs.io/en/latest/index.html) although SqlAlchemy is the standard. For example we would define `models` for our tables: ``` class Project(BaseModel): access_token = pw.TextField() chunk_size = pw.BigIntegerField(constraints=[pw.SQL("DEFAULT...

hardbyte

This service doesn't provide an api for client side filter of p-sig blocks, rather all candidate blocks are uploaded and the server [joins by block name](https://github.com/data61/anonlink-entity-service/blob/1b4a89ba2f639971eb8cf4288c45955a443d9d9b/backend/entityservice/tasks/comparing.py#L182). See the [blocklib tutorial](https://blocklib.readthedocs.io/en/latest/tutorial/tutorial_blocking.html#Generate-Final-Blocks)...

hardbyte

Handle blocking input that doesn't cover all encodings

1

`blocklib` filters out records according to the blocking specification. It warns if all records are not included in a block after using a particular blocking schema, but as someone else...

hardbyte

Improve opentracing

5

Misc issues when trialing jaeger tracing. 1) I've seen the warning "invalid parent span" for the following tasks at various times (not always): - `entityservice.tasks.solver.solver_task`, - `entityservice.tasks.mark_run_complete.mark_run_complete` and - `entityservice.tasks.run.prerun_check`...

hardbyte

Yeah the double up of `models` modules isn't ideal. As you can tell I have avoided the large task of moving off directly using `psycopg2` to using sqlalchemy in the...

wilko77

Add Container Diagrams to docs

2

We should have a high level architecture diagram showing the communication between the various containers.

hardbyte

Create smaller docker images using multi-stage build

3

Instead of including gcc and everything required to compile all our dependencies we could probably build a smaller image which would have a smaller attack surface. https://docs.docker.com/develop/develop-images/multistage-build/#before-multi-stage-builds

hardbyte

That's the file views/project.py. Method project_clks_post needs to support both upload types. The plan is to give the new format the same treatment as the old one. That is: -...

joyceyuu

Configuration of structlog output in logging file

1

See https://github.com/n1analytics/entity-service/pull/349#discussion_r273343179

hardbyte