Compare commits

51 Commits
Author SHA1 Message Date
Joshua Coles df3b84079c Correct account lookup bug
Build and Publish / Build and Test (push) Successful in 7m41s
2025-12-28 12:48:55 +00:00
Joshua Coles d11e4fd0c4 Allow specifying the account ID and fix bug when delivering batched rows
Build and Publish / Build and Test (push) Successful in 9m12s
2025-12-28 12:04:53 +00:00
Joshua Coles a2ba83e6f8 Correct tests 2025-12-28 11:57:29 +00:00
Joshua Coles 0d564ff299 Update deps 2025-12-28 11:57:22 +00:00
joshuacoles bc4aa8242c Fix silly error
Build and Publish / Build and Test (push) Successful in 8m29s
2024-10-31 12:10:57 +00:00
joshuacoles 9d23828345 Revert "Try sccache"
Build and Publish / Build and Test (push) Failing after 6m3s
This reverts commit 6d2150d1b2.
2024-10-31 12:04:08 +00:00
joshuacoles e60ab43abd Add /shortcuts-csv-import
Build and Publish / Build and Test (push) Has been cancelled
2024-10-31 12:03:36 +00:00
joshuacoles 6d2150d1b2 Try sccache
Build and Publish / Build and Test (push) Failing after 3m14s
2024-10-17 09:59:37 +01:00
joshuacoles 7f78d2274c Update debian version
Build and Publish / Build and Test (push) Successful in 2m24s
2024-10-16 13:57:08 +01:00
joshuacoles e739d5ea5b Turns out you just can't use build args in the entrypoint... why... who knows
Build and Publish / Build and Test (push) Successful in 2m19s
2024-10-16 13:39:46 +01:00
joshuacoles 09209be500 Take 4
Build and Publish / Build and Test (push) Failing after 2m14s
2024-10-16 13:13:45 +01:00
joshuacoles c86a79b46e Take 3
Build and Publish / Build and Test (push) Failing after 2m12s
2024-10-16 13:04:59 +01:00
joshuacoles b65fad16e4 Take 2
Build and Publish / Build and Test (push) Successful in 2m38s
2024-10-16 09:40:23 +01:00
joshuacoles a60d0effff Remove quotes around binary name
Build and Publish / Build and Test (push) Successful in 11m0s
Maybe this will work?
2024-10-15 22:46:28 +01:00
joshuacoles c230aef034 Correct target name from gnu to musl in Dockerfile
Build and Publish / Build and Test (push) Successful in 10m32s
2024-10-15 21:42:36 +01:00
joshuacoles ca0df97e73 Slim down batch size to fit in ingestion
Build and Publish / Build and Test (push) Waiting to run
2024-10-15 21:40:11 +01:00
joshuacoles 5c3f734bbc Support account IDs 2024-10-15 21:35:55 +01:00
joshuacoles e3ed72c9b0 Create monzo and flex accounts 2024-10-15 21:17:39 +01:00
joshuacoles 3c3b6dc4e6 Add some flex migrations 2024-10-15 21:03:40 +01:00
joshuacoles c3796720b7 Bump sea-orm version 2024-10-15 20:50:06 +01:00
joshuacoles 35fd2b90d2 Attempt to fix build with ignore rules
Build and Publish / Build and Test (push) Failing after 11m55s
2024-09-08 17:51:49 +01:00
joshuacoles 4b2f0f3bf7 Update build.yml file
Build and Publish / Build and Test (push) Failing after 8m7s
2024-09-08 14:35:28 +01:00
joshuacoles 29fe8bee39 Disable tests as they aren't designed to run in CI yet
Rust CI / Build and Test (push) Failing after 7m14s
2024-08-09 12:31:01 +01:00
joshuacoles 49a1700706 Move docker setup up to try make the tests pass
Rust CI / Build and Test (push) Failing after 34m17s
2024-08-09 11:53:33 +01:00
joshuacoles 04b58d3075 Fix binary name
Rust CI / Build and Test (push) Failing after 44m28s
2024-08-09 10:28:21 +01:00
joshuacoles 076a573711 Bump versions once more and move to specifying only X.Y versions to make this easier
Rust CI / Build and Test (push) Has been cancelled
2024-08-09 10:07:51 +01:00
joshuacoles af0588c5ef Replace the Docker build with the one from toggl-bridge as it's better
Rust CI / Build and Test (push) Has been cancelled
2024-08-09 09:57:48 +01:00
joshuacoles 21a63a10a4 Fix some clippy lints 2024-08-09 09:56:23 +01:00
joshuacoles 08766dc0e0 Stash 2024-08-09 09:43:10 +01:00
joshuacoles bf47520d31 Try to make the cache work, take N
Build and Publish Docker Container / build (push) Successful in 16m35s
2024-06-04 08:42:46 +01:00
joshuacoles fc1cea32b5 Pin docker/build-push-action@v5 and add ntfy.sh action 2024-06-03 21:12:02 +01:00
joshuacoles 92462bd316 Try to use GHA cache instead
Build and Publish Docker Container / build (push) Failing after 13m16s
2024-06-03 20:46:08 +01:00
joshuacoles f0b0cb1567 Start tagging with SHA as well as latest 2024-06-03 20:45:39 +01:00
joshuacoles 8478fa0b38 Try manually cache build layers
Build and Publish Docker Container / build (push) Has been cancelled
2024-06-03 20:38:54 +01:00
joshuacoles 3b2c1aeda0 Add some documentation to the CLI and the csv ingestion route
Build and Publish Docker Container / build (push) Has been cancelled
2024-06-03 20:26:31 +01:00
joshuacoles b37273cfbe Fix Dockerfile.cache glibc version
Build and Publish Docker Container / build (push) Successful in 14m11s
2024-06-03 20:14:41 +01:00
joshuacoles f344d69419 Fix Dockerfile.cache 2024-06-03 20:09:46 +01:00
joshuacoles 901aba9c7f Fix Dockerfile.cache 2024-06-03 20:09:26 +01:00
joshuacoles d7d7fa9718 Investigate cargo-chef to cache things
Build and Publish Docker Container / build (push) Has been cancelled
2024-06-03 20:09:00 +01:00
joshuacoles b8c1faced2 Update Dockerfile to match new command structure 2024-06-03 20:03:43 +01:00
joshuacoles b563bbd02c Add a little more CLI structure for local runs 2024-06-03 19:59:22 +01:00
joshuacoles 7fd85550ea Try to improve caching?
Build and Publish Docker Container / build (push) Successful in 9m41s
2024-06-03 19:19:50 +01:00
joshuacoles 97f57803e5 Expose more information in logging and improve error handling
Build and Publish Docker Container / build (push) Successful in 9m28s
2024-06-03 19:04:44 +01:00
joshuacoles 046ce44d23 rustfmt
Build and Publish Docker Container / build (push) Successful in 10m2s
2024-06-03 18:41:32 +01:00
joshuacoles f70d844ff3 Fix ORM failing to handle empty lists... 2024-06-03 18:41:21 +01:00
joshuacoles 61de38b9bf Add monzo-row parsing testcases
Build and Publish Docker Container / build (push) Successful in 9m19s
2024-06-03 17:41:43 +01:00
joshuacoles 6c5d3910dc Make title optional in db 2024-06-03 17:34:08 +01:00
joshuacoles 4bb9f2813d Move transaction whittling out of the database
Build and Publish Docker Container / build (push) Successful in 9m56s
2024-06-03 12:12:45 +01:00
joshuacoles 3df05b2d9c Add identity_hash column
Build and Publish Docker Container / build (push) Failing after 9m27s
2024-05-29 21:08:51 +01:00
joshuacoles f19f861297 Introduce MonzoRow as an intermediary between json/csv and the database entities. This will allow for hashing to perform identity checks 2024-05-29 20:49:48 +01:00
joshuacoles 89d0d12e26 Make SQLx logging debug level and add tower-http traces
Build and Publish Docker Container / build (push) Successful in 10m20s
2024-05-29 20:12:40 +01:00
29 changed files with 2810 additions and 1459 deletions
+1
View File
@@ -28,6 +28,7 @@
**/values.dev.yaml **/values.dev.yaml
/bin /bin
/target /target
!/target/**/monzo-ingestion
/.idea /.idea
LICENSE LICENSE
README.md README.md
+1 -1
View File
@@ -1 +1 @@
DATABASE_URL=postgres://postgres@localhost/logos_neu DATABASE_URL=postgres://postgres@localhost/monzo_development
+82 -13
View File
@@ -1,33 +1,102 @@
name: Build and Publish Docker Container name: Build and Publish
on: on:
push: push:
branches: branches: [ main ]
- main
env:
CARGO_TERM_COLOR: always
RUST_BINARY_NAME: monzo-ingestion
jobs: jobs:
build: build:
name: Build and Test
runs-on: ubuntu-latest runs-on: ubuntu-latest
container: container: catthehacker/ubuntu:act-latest
image: catthehacker/ubuntu:act-latest
steps: steps:
- name: Checkout code - uses: actions/checkout@v3
uses: actions/checkout@v2
- name: Install Rust
uses: actions-rs/toolchain@v1
with:
toolchain: stable
profile: minimal
override: true
components: rustfmt, clippy
- name: Add ARM64 target
run: rustup target add aarch64-unknown-linux-musl
- name: Install ARM64 toolchain
run: |
apt-get update
apt-get install -y gcc-aarch64-linux-gnu build-essential musl-tools
- name: Cache dependencies
uses: actions/cache@v3
with:
path: |
~/.cargo
target/
key: "${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}"
restore-keys: |
${{ runner.os }}-cargo-
- name: Build (x86_64)
uses: actions-rs/cargo@v1
with:
command: build
args: --release --all-features
- name: Build (ARM64)
uses: actions-rs/cargo@v1
with:
command: build
args: --release --all-features --target aarch64-unknown-linux-musl
env:
CARGO_TARGET_AARCH64_UNKNOWN_LINUX_MUSL_LINKER: aarch64-linux-gnu-gcc
- name: Set up QEMU
uses: docker/setup-qemu-action@v2
- name: Set up Docker Buildx - name: Set up Docker Buildx
uses: docker/setup-buildx-action@v1 uses: docker/setup-buildx-action@v2
- name: Login to Docker - name: Login to DockerHub
uses: docker/login-action@v1 uses: docker/login-action@v2
with: with:
registry: git.joshuacoles.me registry: git.joshuacoles.me
username: ${{ secrets.DOCKER_USERNAME }} username: ${{ secrets.DOCKER_USERNAME }}
password: ${{ secrets.DOCKER_PASSWORD }} password: ${{ secrets.DOCKER_PASSWORD }}
- name: Build and Push Docker image - name: Upload artifacts
uses: docker/build-push-action@master uses: actions/upload-artifact@v3
with:
name: binaries
path: |
target/release/${{ env.RUST_BINARY_NAME }}
target/aarch64-unknown-linux-musl/release/${{ env.RUST_BINARY_NAME }}
# # The target directory is kept in the .dockerignore file, so allow these to be copied in we need to move them to a
# # new directory.
# - run: mv target docker-binaries
- name: Build and push multi-arch Docker image
uses: docker/build-push-action@v4
with: with:
context: . context: .
file: ./Dockerfile
platforms: linux/amd64,linux/arm64
push: true push: true
tags: git.joshuacoles.me/personal/monzo-ingestion:latest tags: git.joshuacoles.me/${{ github.repository }}:${{ github.sha }},git.joshuacoles.me/${{ github.repository }}:latest
build-args: |
BINARY_NAME=${{ env.RUST_BINARY_NAME }}
- uses: robiningelbrecht/ntfy-action@v1.0.0
name: Notify via ntfy.sh
if: always()
with:
url: ${{ secrets.NTFY_URL }}
topic: ${{ secrets.NTFY_TOPIC }}
job_status: ${{ job.status }}
Generated
+1653 -1093
View File
File diff suppressed because it is too large Load Diff
+17 -14
View File
@@ -7,28 +7,31 @@ edition = "2021"
entity = { path = "entity" } entity = { path = "entity" }
migration = { path = "migration" } migration = { path = "migration" }
axum = { version = "0.7.5", features = ["multipart"] } axum = { version = "0.8.8", features = ["multipart"] }
tokio = { version = "1.37.0", features = ["full"] } tokio = { version = "1.48.0", features = ["full"] }
sea-orm = { version = "1.0.0-rc.4", features = [ sea-orm = { version = "1.1.0", features = [
"sqlx-postgres", "sqlx-postgres",
"runtime-tokio-rustls", "runtime-tokio-rustls",
"macros" "macros"
] } ] }
serde = { version = "1.0.203", features = ["derive"] } serde = { version = "1.0", features = ["derive"] }
serde_json = "1.0.117" serde_json = "1.0"
tracing-subscriber = "0.3.18" tracing-subscriber = "0.3.18"
tracing = "0.1.40" tracing = "0.1.40"
anyhow = "1.0.86" anyhow = { version = "1.0", features = ["backtrace"] }
thiserror = "1.0.61" thiserror = "2.0"
http = "1.1.0" http = "1.1"
chrono = { version = "0.4.38", features = ["serde"] } chrono = { version = "0.4", features = ["serde"] }
num-traits = "0.2.19" num-traits = "0.2"
csv = "1.3.0" csv = "1.3.0"
clap = "4.5.4" clap = "4.5"
testcontainers = "0.17.0" testcontainers = "0.26"
testcontainers-modules = { version = "0.5.0", features = ["postgres"] } testcontainers-modules = { version = "0.14", features = ["postgres"] }
sqlx = { version = "0.7.4", features = ["postgres"] } sqlx = { version = "0.8", features = ["postgres"] }
tower-http = { version = "0.6", features = ["trace"] }
bytes = "1.7"
once_cell = "1.19"
[workspace] [workspace]
members = [".", "migration", "entity"] members = [".", "migration", "entity"]
+12 -76
View File
@@ -1,78 +1,14 @@
# syntax=docker/dockerfile:1 FROM busybox AS platform_determiner
ARG TARGETPLATFORM
ARG BINARY_NAME
# Comments are provided throughout this file to help you get started. COPY /target /target
# If you need more help, visit the Dockerfile reference guide at RUN case "$TARGETPLATFORM" in \
# https://docs.docker.com/engine/reference/builder/ "linux/amd64") BINARY_PATH="/target/release/${BINARY_NAME}" ;; \
"linux/arm64") BINARY_PATH="/target/aarch64-unknown-linux-musl/release/${BINARY_NAME}" ;; \
*) exit 1 ;; \
esac && mv "$BINARY_PATH" "/usr/bin/monzo-ingestion" && chmod +x "/usr/bin/monzo-ingestion"
################################################################################ FROM --platform=$TARGETPLATFORM debian:bookworm-slim
# Create a stage for building the application. COPY --from=platform_determiner /usr/bin/monzo-ingestion /usr/local/bin/monzo-ingestion
ENTRYPOINT ["/usr/local/bin/monzo-ingestion"]
ARG RUST_VERSION=1.76.0
ARG APP_NAME=monzo-ingestion
FROM rust:${RUST_VERSION}-slim-bullseye AS build
ARG APP_NAME
WORKDIR /app
# Build the application.
# Leverage a cache mount to /usr/local/cargo/registry/
# for downloaded dependencies and a cache mount to /app/target/ for
# compiled dependencies which will speed up subsequent builds.
# Leverage a bind mount to the src directory to avoid having to copy the
# source code into the container. Once built, copy the executable to an
# output directory before the cache mounted /app/target is unmounted.
RUN --mount=type=bind,source=src,target=src \
--mount=type=bind,source=entity,target=entity \
--mount=type=bind,source=migration,target=migration \
--mount=type=bind,source=Cargo.toml,target=Cargo.toml \
--mount=type=bind,source=Cargo.lock,target=Cargo.lock \
--mount=type=cache,target=/app/target/ \
--mount=type=cache,target=/usr/local/cargo/registry/ \
<<EOF
set -e
cargo build --locked --release
cp ./target/release/$APP_NAME /bin/server
EOF
################################################################################
# Create a new stage for running the application that contains the minimal
# runtime dependencies for the application. This often uses a different base
# image from the build stage where the necessary files are copied from the build
# stage.
#
# The example below uses the debian bullseye image as the foundation for running the app.
# By specifying the "bullseye-slim" tag, it will also use whatever happens to be the
# most recent version of that tag when you build your Dockerfile. If
# reproducability is important, consider using a digest
# (e.g., debian@sha256:ac707220fbd7b67fc19b112cee8170b41a9e97f703f588b2cdbbcdcecdd8af57).
FROM debian:bullseye-slim AS final
RUN set -ex; \
apt-get update && \
apt-get -y install --no-install-recommends \
ca-certificates curl && \
rm -rf /var/lib/apt/lists/*
# Create a non-privileged user that the app will run under.
# See https://docs.docker.com/develop/develop-images/dockerfile_best-practices/#user
ARG UID=10001
RUN adduser \
--disabled-password \
--gecos "" \
--home "/nonexistent" \
--shell "/sbin/nologin" \
--no-create-home \
--uid "${UID}" \
appuser
USER appuser
# Copy the executable from the "build" stage.
COPY --from=build /bin/server /bin/
# Expose the port that the application listens on.
EXPOSE 3000
HEALTHCHECK --interval=5s --timeout=3s --retries=3 \
CMD curl -f http://localhost:3000/health || exit 1
# What the container should run when it is started.
CMD ["/bin/server"]
+1 -1
View File
@@ -9,5 +9,5 @@ name = "entity"
path = "src/lib.rs" path = "src/lib.rs"
[dependencies] [dependencies]
sea-orm = { version = "1.0.0-rc.4" } sea-orm = { version = "1.1.0" }
serde = { version = "1.0.203", features = ["derive"] } serde = { version = "1.0.203", features = ["derive"] }
+26
View File
@@ -0,0 +1,26 @@
//! `SeaORM` Entity, @generated by sea-orm-codegen 1.1.0
use sea_orm::entity::prelude::*;
use serde::{Deserialize, Serialize};
#[derive(Clone, Debug, PartialEq, DeriveEntityModel, Eq, Serialize, Deserialize)]
#[sea_orm(table_name = "account")]
pub struct Model {
#[sea_orm(primary_key)]
pub id: i32,
pub name: String,
}
#[derive(Copy, Clone, Debug, EnumIter, DeriveRelation)]
pub enum Relation {
#[sea_orm(has_many = "super::transaction::Entity")]
Transaction,
}
impl Related<super::transaction::Entity> for Entity {
fn to() -> RelationDef {
Relation::Transaction.def()
}
}
impl ActiveModelBehavior for ActiveModel {}
+17 -2
View File
@@ -1,4 +1,4 @@
//! `SeaORM` Entity. Generated by sea-orm-codegen 0.12.2 //! `SeaORM` Entity, @generated by sea-orm-codegen 1.1.0
use sea_orm::entity::prelude::*; use sea_orm::entity::prelude::*;
use serde::{Deserialize, Serialize}; use serde::{Deserialize, Serialize};
@@ -14,6 +14,21 @@ pub struct Model {
} }
#[derive(Copy, Clone, Debug, EnumIter, DeriveRelation)] #[derive(Copy, Clone, Debug, EnumIter, DeriveRelation)]
pub enum Relation {} pub enum Relation {
#[sea_orm(
belongs_to = "super::transaction::Entity",
from = "Column::TransactionId",
to = "super::transaction::Column::Id",
on_update = "NoAction",
on_delete = "Cascade"
)]
Transaction,
}
impl Related<super::transaction::Entity> for Entity {
fn to() -> RelationDef {
Relation::Transaction.def()
}
}
impl ActiveModelBehavior for ActiveModel {} impl ActiveModelBehavior for ActiveModel {}
+2 -1
View File
@@ -1,6 +1,7 @@
//! `SeaORM` Entity. Generated by sea-orm-codegen 0.12.2 //! `SeaORM` Entity, @generated by sea-orm-codegen 1.1.0
pub mod prelude; pub mod prelude;
pub mod account;
pub mod expenditure; pub mod expenditure;
pub mod transaction; pub mod transaction;
+2 -1
View File
@@ -1,4 +1,5 @@
//! `SeaORM` Entity. Generated by sea-orm-codegen 0.12.2 //! `SeaORM` Entity, @generated by sea-orm-codegen 1.1.0
pub use super::account::Entity as Account;
pub use super::expenditure::Entity as Expenditure; pub use super::expenditure::Entity as Expenditure;
pub use super::transaction::Entity as Transaction; pub use super::transaction::Entity as Transaction;
+29 -3
View File
@@ -1,4 +1,4 @@
//! `SeaORM` Entity. Generated by sea-orm-codegen 0.12.2 //! `SeaORM` Entity, @generated by sea-orm-codegen 1.1.0
use sea_orm::entity::prelude::*; use sea_orm::entity::prelude::*;
use serde::{Deserialize, Serialize}; use serde::{Deserialize, Serialize};
@@ -11,14 +11,40 @@ pub struct Model {
pub transaction_type: String, pub transaction_type: String,
pub total_amount: Decimal, pub total_amount: Decimal,
pub timestamp: DateTime, pub timestamp: DateTime,
pub title: String, pub title: Option<String>,
pub emoji: Option<String>, pub emoji: Option<String>,
pub notes: Option<String>, pub notes: Option<String>,
pub receipt: Option<String>, pub receipt: Option<String>,
pub description: Option<String>, pub description: Option<String>,
#[sea_orm(unique)]
pub identity_hash: Option<i64>,
pub account_id: Option<i32>,
} }
#[derive(Copy, Clone, Debug, EnumIter, DeriveRelation)] #[derive(Copy, Clone, Debug, EnumIter, DeriveRelation)]
pub enum Relation {} pub enum Relation {
#[sea_orm(
belongs_to = "super::account::Entity",
from = "Column::AccountId",
to = "super::account::Column::Id",
on_update = "NoAction",
on_delete = "Cascade"
)]
Account,
#[sea_orm(has_many = "super::expenditure::Entity")]
Expenditure,
}
impl Related<super::account::Entity> for Entity {
fn to() -> RelationDef {
Relation::Account.def()
}
}
impl Related<super::expenditure::Entity> for Entity {
fn to() -> RelationDef {
Relation::Expenditure.def()
}
}
impl ActiveModelBehavior for ActiveModel {} impl ActiveModelBehavior for ActiveModel {}
+3
View File
@@ -0,0 +1,3 @@
Transaction ID,Date,Time,Type,Name,Emoji,Category,Amount,Currency,Local amount,Local currency,Notes and #tags,Address,Receipt,Description,Category split
tx_0000AhmcbFKVlkAFbeZuAk,11/05/2024,13:41:47,Pot transfer,Savings Pot,,Savings,-1.4,GBP,-1.4,GBP,,,,Round up,
tx_0000AhmgdZ0FzkVZou77Ev,11/05/2024,14:27:01,Card payment,D,💊,C,-100.42,GBP,-100.42,GBP,,B,,A,"Groceries:15.00,Personal care:10.00"
1 Transaction ID Date Time Type Name Emoji Category Amount Currency Local amount Local currency Notes and #tags Address Receipt Description Category split
2 tx_0000AhmcbFKVlkAFbeZuAk 11/05/2024 13:41:47 Pot transfer Savings Pot Savings -1.4 GBP -1.4 GBP Round up
3 tx_0000AhmgdZ0FzkVZou77Ev 11/05/2024 14:27:01 Card payment D 💊 C -100.42 GBP -100.42 GBP B A Groceries:15.00,Personal care:10.00
+38
View File
@@ -0,0 +1,38 @@
[
[
"tx_0000AhmcbFKVlkAFbeZuAk",
"2024-05-11T00:00:00.000Z",
"1899-12-30T13:41:47.000Z",
"Pot transfer",
"Savings Pot",
"",
"Savings",
-1.4,
"GBP",
-1.4,
"GBP",
"",
"",
"",
"Round up",
""
],
[
"tx_0000AhmgdZ0FzkVZou77Ev",
"2024-05-11T00:00:00.000Z",
"1899-12-30T14:27:01.000Z",
"Card payment",
"D",
"💊",
"C",
-100.42,
"GBP",
-100.42,
"GBP",
"",
"B",
"",
"A",
"Groceries:15.00,Personal care:10.00"
]
]
Executable
+12
View File
@@ -0,0 +1,12 @@
#!/usr/bin/env just --justfile
generate-migration:
sea-orm-cli migrate generate $1
migrate:
sea-orm-cli migrate up
generate-entities:
sea-orm-cli generate entity --with-serde both \
-l -o entity/src \
--ignore-tables monzo_ingestion_seaql_migrations
+1 -1
View File
@@ -12,7 +12,7 @@ path = "src/lib.rs"
async-std = { version = "1", features = ["attributes", "tokio1"] } async-std = { version = "1", features = ["attributes", "tokio1"] }
[dependencies.sea-orm-migration] [dependencies.sea-orm-migration]
version = "1.0.0-rc.4" version = "1.1.0"
features = [ features = [
"runtime-tokio-rustls", # `ASYNC_RUNTIME` feature "runtime-tokio-rustls", # `ASYNC_RUNTIME` feature
"sqlx-postgres", # `DATABASE_DRIVER` feature "sqlx-postgres", # `DATABASE_DRIVER` feature
+13 -1
View File
@@ -1,6 +1,11 @@
pub use sea_orm_migration::prelude::*; pub use sea_orm_migration::prelude::*;
pub mod m20230904_141851_create_monzo_tables; pub mod m20230904_141851_create_monzo_tables;
mod m20240529_195030_add_transaction_identity_hash;
mod m20240603_162500_make_title_optional;
mod m20241015_195220_add_account_to_transactions;
mod m20241015_200222_add_expenditure_transaction_fk;
mod m20241015_200652_add_accounts;
pub struct Migrator; pub struct Migrator;
@@ -11,6 +16,13 @@ impl MigratorTrait for Migrator {
} }
fn migrations() -> Vec<Box<dyn MigrationTrait>> { fn migrations() -> Vec<Box<dyn MigrationTrait>> {
vec![Box::new(m20230904_141851_create_monzo_tables::Migration)] vec![
Box::new(m20230904_141851_create_monzo_tables::Migration),
Box::new(m20240529_195030_add_transaction_identity_hash::Migration),
Box::new(m20240603_162500_make_title_optional::Migration),
Box::new(m20241015_195220_add_account_to_transactions::Migration),
Box::new(m20241015_200222_add_expenditure_transaction_fk::Migration),
Box::new(m20241015_200652_add_accounts::Migration),
]
} }
} }
@@ -0,0 +1,39 @@
use sea_orm_migration::prelude::*;
#[derive(DeriveMigrationName)]
pub struct Migration;
#[async_trait::async_trait]
impl MigrationTrait for Migration {
async fn up(&self, manager: &SchemaManager) -> Result<(), DbErr> {
manager
.alter_table(
TableAlterStatement::new()
.table(Transaction::Table)
.add_column(
ColumnDef::new(Transaction::IdentityHash)
.big_integer()
.unique_key(),
)
.to_owned(),
)
.await
}
async fn down(&self, manager: &SchemaManager) -> Result<(), DbErr> {
manager
.alter_table(
TableAlterStatement::new()
.table(Transaction::Table)
.drop_column(Transaction::IdentityHash)
.to_owned(),
)
.await
}
}
#[derive(DeriveIden)]
enum Transaction {
Table,
IdentityHash,
}
@@ -0,0 +1,61 @@
use sea_orm_migration::prelude::*;
#[derive(DeriveMigrationName)]
pub struct Migration;
#[async_trait::async_trait]
impl MigrationTrait for Migration {
async fn up(&self, manager: &SchemaManager) -> Result<(), DbErr> {
manager
.alter_table(
TableAlterStatement::new()
.table(Transaction::Table)
.modify_column(ColumnDef::new(Transaction::Title).string().null())
.to_owned(),
)
.await?;
// Set all empty string titles to null
manager
.get_connection()
.execute_unprepared(
r#"
update transaction
set title = null
where title = ''
"#,
)
.await?;
Ok(())
}
async fn down(&self, manager: &SchemaManager) -> Result<(), DbErr> {
// Set all null titles to empty string when reverting
manager
.get_connection()
.execute_unprepared(
r#"
update transaction
set title = ''
where title is null
"#,
)
.await?;
manager
.alter_table(
TableAlterStatement::new()
.table(Transaction::Table)
.modify_column(ColumnDef::new(Transaction::Title).string().not_null())
.to_owned(),
)
.await
}
}
#[derive(DeriveIden)]
enum Transaction {
Table,
Title,
}
@@ -0,0 +1,76 @@
use sea_orm_migration::{prelude::*, schema::*};
#[derive(DeriveMigrationName)]
pub struct Migration;
#[async_trait::async_trait]
impl MigrationTrait for Migration {
async fn up(&self, manager: &SchemaManager) -> Result<(), DbErr> {
manager
.create_table(
Table::create()
.table(Account::Table)
.if_not_exists()
.col(
ColumnDef::new(Account::Id)
.integer()
.auto_increment()
.not_null()
.primary_key(),
)
.col(
ColumnDef::new(Account::Name)
.string()
.not_null(),
)
.to_owned(),
)
.await?;
manager.alter_table(
TableAlterStatement::new()
.table(Transaction::Table)
.add_column(ColumnDef::new(Transaction::AccountId).integer())
.to_owned(),
).await?;
manager
.create_foreign_key(
ForeignKey::create()
.name("fk_transaction_account_id")
.from(Transaction::Table, Transaction::AccountId)
.to(Account::Table, Account::Id)
.on_delete(ForeignKeyAction::Cascade)
.to_owned(),
)
.await?;
Ok(())
}
async fn down(&self, manager: &SchemaManager) -> Result<(), DbErr> {
manager.alter_table(
TableAlterStatement::new()
.table(Transaction::Table)
.drop_column(Transaction::AccountId)
.to_owned(),
).await?;
manager.drop_table(Table::drop().table(Account::Table).to_owned()).await?;
Ok(())
}
}
#[derive(DeriveIden)]
enum Account {
Table,
Id,
Name,
}
#[derive(DeriveIden)]
enum Transaction {
Table,
AccountId,
}
@@ -0,0 +1,40 @@
use sea_orm_migration::{prelude::*, schema::*};
#[derive(DeriveMigrationName)]
pub struct Migration;
#[async_trait::async_trait]
impl MigrationTrait for Migration {
async fn up(&self, manager: &SchemaManager) -> Result<(), DbErr> {
manager
.create_foreign_key(
ForeignKey::create()
.name("fk_expenditure_transaction_id")
.from(Expenditure::Table, Expenditure::TransactionId)
.to(Transaction::Table, Transaction::Id)
.on_delete(ForeignKeyAction::Cascade)
.to_owned(),
)
.await?;
Ok(())
}
async fn down(&self, manager: &SchemaManager) -> Result<(), DbErr> {
manager.drop_foreign_key(ForeignKey::drop().name("fk_expenditure_transaction_id").to_owned()).await?;
Ok(())
}
}
#[derive(DeriveIden)]
enum Transaction {
Table,
Id,
}
#[derive(DeriveIden)]
enum Expenditure {
Table,
TransactionId,
}
@@ -0,0 +1,53 @@
use sea_orm_migration::{prelude::*, schema::*};
use crate::sea_orm::sqlx;
#[derive(DeriveMigrationName)]
pub struct Migration;
#[async_trait::async_trait]
impl MigrationTrait for Migration {
async fn up(&self, manager: &SchemaManager) -> Result<(), DbErr> {
manager.exec_stmt(
InsertStatement::new()
.into_table(Account::Table)
.columns(vec![Account::Name])
.values_panic(["Monzo".into()])
.values_panic(["Flex".into()])
.to_owned()
).await?;
let id: i32 = manager.get_connection().query_one(
sea_orm::Statement::from_string(
manager.get_database_backend(),
"SELECT id FROM account WHERE name = 'Monzo'",
)
).await?.expect("Monzo account not found").try_get_by_index(0)?;
manager.exec_stmt(
UpdateStatement::new()
.table(Transaction::Table)
.values([(Transaction::AccountId, id.into())])
.to_owned()
).await?;
Ok(())
}
async fn down(&self, manager: &SchemaManager) -> Result<(), DbErr> {
// This is a data creation migration, so we do not roll back the migration
Ok(())
}
}
#[derive(DeriveIden)]
enum Account {
Table,
Id,
Name,
}
#[derive(DeriveIden)]
enum Transaction {
Table,
AccountId,
}
+13 -5
View File
@@ -6,25 +6,33 @@ use tracing::log::error;
#[derive(thiserror::Error, Debug)] #[derive(thiserror::Error, Debug)]
pub enum AppError { pub enum AppError {
/// SeaORM error, separated for ease of use allowing us to `?` db operations. /// SeaORM error, separated for ease of use allowing us to `?` db operations.
#[error("Internal error")] #[error("Database error: {0}")]
DbError(#[from] DbErr), DbError(#[from] DbErr),
#[error("Invalid request {0}")] #[error("Invalid request {0}")]
BadRequest(anyhow::Error), BadRequest(anyhow::Error),
/// Catch all for error we don't care to expose publicly. /// Catch all for error we don't care to expose publicly.
#[error("Internal error")] #[error("An error occurred: {0}")]
Anyhow(#[from] anyhow::Error), Anyhow(#[from] anyhow::Error),
} }
impl AppError {
fn to_response_string(&self) -> String {
match self {
AppError::BadRequest(e) => e.to_string(),
_ => "Internal server error".to_string(),
}
}
}
impl IntoResponse for AppError { impl IntoResponse for AppError {
fn into_response(self) -> Response { fn into_response(self) -> Response {
error!("Internal server error: {self:?}");
let status_code = match self { let status_code = match self {
AppError::BadRequest(_) => StatusCode::BAD_REQUEST,
_ => StatusCode::INTERNAL_SERVER_ERROR, _ => StatusCode::INTERNAL_SERVER_ERROR,
}; };
(status_code, self.to_string()).into_response() (status_code, self.to_response_string()).into_response()
} }
} }
+241 -71
View File
@@ -1,34 +1,75 @@
use crate::error::AppError; use crate::error::AppError;
use crate::ingestion::ingestion_logic::MonzoRow;
use anyhow::anyhow; use anyhow::anyhow;
use entity::{expenditure, transaction}; use entity::{expenditure, transaction};
use sea_orm::sea_query::{OnConflict, PostgresQueryBuilder}; use sea_orm::sea_query::OnConflict;
use sea_orm::{ColumnTrait, DatabaseConnection, EntityTrait, Iterable, TransactionTrait};
use sea_orm::{ use sea_orm::{
ConnectionTrait, DatabaseBackend, DatabaseTransaction, DbErr, QueryFilter, QueryTrait, ColumnTrait, DatabaseConnection, DbErr, EntityTrait, Iterable, QuerySelect, TransactionTrait,
Statement,
}; };
use sea_orm::{ConnectionTrait, DatabaseTransaction, QueryFilter};
pub struct Insertion { pub struct Insertion {
pub transaction: transaction::ActiveModel, pub transaction: transaction::ActiveModel,
pub contained_expenditures: Vec<expenditure::ActiveModel>, pub contained_expenditures: Vec<expenditure::ActiveModel>,
pub identity_hash: i64,
}
// Note while this is more efficient in db calls, it does bind together the entire group.
// We employ a batching process for now to try balance speed and failure rate, but it is worth
// trying to move failures earlier and improve reporting.
pub async fn insert(
db: &DatabaseConnection,
monzo_rows: Vec<MonzoRow>,
account_id: i32,
) -> Result<Vec<String>, AppError> {
let mut new_transaction_ids = Vec::new();
let insertions = monzo_rows
.into_iter()
.map(|row| MonzoRow::into_insertion(row, account_id))
.collect::<Result<Vec<_>, _>>()?;
for insertions in insertions.chunks(200) {
let (new_or_updated_insertions, inserted_transaction_ids) =
whittle_insertions(insertions, db).await?;
if new_or_updated_insertions.is_empty() {
continue;
}
let tx = db.begin().await?;
update_transactions(&tx, &new_or_updated_insertions).await?;
update_expenditures(&tx, &new_or_updated_insertions, &inserted_transaction_ids).await?;
tx.commit().await?;
// We wait until the transaction is committed before adding the new transaction ids to the
// list to avoid issues with the transaction being rolled back.
new_transaction_ids.extend(inserted_transaction_ids);
}
// Notify the new transactions once everything is committed.
notify_new_transactions(db, &new_transaction_ids).await?;
Ok(new_transaction_ids)
} }
async fn update_expenditures( async fn update_expenditures(
tx: &DatabaseTransaction, tx: &DatabaseTransaction,
insertions: &[Insertion], new_or_updated_insertions: &[&Insertion],
) -> Result<(), DbErr> { inserted_transaction_ids: &[String],
) -> Result<(), AppError> {
if new_or_updated_insertions.is_empty() {
return Ok(());
}
// Expenditures can change as we re-categorise them, so we delete all the old ones and // Expenditures can change as we re-categorise them, so we delete all the old ones and
// insert an entirely new set to ensure we don't end up leaving old ones around. // insert an entirely new set to ensure we don't end up leaving old ones around.
expenditure::Entity::delete_many() expenditure::Entity::delete_many()
.filter( .filter(expenditure::Column::TransactionId.is_in(inserted_transaction_ids))
expenditure::Column::TransactionId
.is_in(insertions.iter().map(|i| i.transaction.id.as_ref())),
)
.exec(tx) .exec(tx)
.await?; .await?;
expenditure::Entity::insert_many( expenditure::Entity::insert_many(
insertions new_or_updated_insertions
.iter() .iter()
.flat_map(|i| &i.contained_expenditures) .flat_map(|i| &i.contained_expenditures)
.cloned(), .cloned(),
@@ -43,62 +84,65 @@ async fn update_expenditures(
) )
.exec(tx) .exec(tx)
.await?; .await?;
Ok(()) Ok(())
} }
// Note while this is more efficient in db calls, it does bind together the entire group. async fn update_transactions(
// We employ a batching process for now to try balance speed and failure rate, but it is worth tx: &DatabaseTransaction,
// trying to move failures earlier and improve reporting. new_or_updated_insertions: &[&Insertion],
pub async fn insert( ) -> Result<(), DbErr> {
db: &DatabaseConnection, if new_or_updated_insertions.is_empty() {
insertions: Vec<Insertion>, return Ok(());
) -> Result<Vec<String>, AppError> {
let mut new_transaction_ids = Vec::new();
for insertions in insertions.chunks(400) {
let tx = db.begin().await?;
let inserted_transaction_ids = update_transactions(insertions, &tx).await?;
update_expenditures(&tx, &insertions).await?;
tx.commit().await?;
// We wait until the transaction is committed before adding the new transaction ids to the
// list to avoid issues with the transaction being rolled back.
new_transaction_ids.extend(inserted_transaction_ids);
} }
// Notify the new transactions once everything is committed. let transactions = new_or_updated_insertions
notify_new_transactions(db, &new_transaction_ids).await?; .iter()
.map(|i| &i.transaction)
.cloned();
Ok(new_transaction_ids) transaction::Entity::insert_many(transactions)
}
async fn update_transactions(
insertions: &[Insertion],
tx: &DatabaseTransaction,
) -> Result<Vec<String>, AppError> {
let insert =
transaction::Entity::insert_many(insertions.iter().map(|i| &i.transaction).cloned())
.on_conflict( .on_conflict(
OnConflict::column(transaction::Column::Id) OnConflict::column(transaction::Column::Id)
.update_columns(transaction::Column::iter()) .update_columns(transaction::Column::iter())
.to_owned(), .to_owned(),
) )
.into_query() .exec(tx)
.returning_col(transaction::Column::Id) .await?;
.build(PostgresQueryBuilder); Ok(())
}
let inserted_transaction_ids = tx async fn whittle_insertions<'a>(
.query_all(Statement::from_sql_and_values( insertions: &'a [Insertion],
DatabaseBackend::Postgres, tx: &DatabaseConnection,
insert.0, ) -> Result<(Vec<&'a Insertion>, Vec<String>), AppError> {
insert.1, let existing_hashes = transaction::Entity::find()
)) .select_only()
.await? .columns([transaction::Column::IdentityHash])
.filter(transaction::Column::IdentityHash.is_not_null())
.into_tuple::<(i64,)>()
.all(tx)
.await?;
tracing::debug!("Found existing entries: {existing_hashes:?}");
// We will only update those where the hash is different to avoid unnecessary updates and
// notifications.
let new_or_updated_insertions = insertions
.iter() .iter()
.map(|r| r.try_get_by("id")) .filter(|i| {
.collect::<Result<Vec<String>, _>>()?; let hash = i.identity_hash;
Ok(inserted_transaction_ids) !existing_hashes
.iter()
.any(|(existing_hash,)| *existing_hash == hash)
})
.collect::<Vec<_>>();
let inserted_transaction_ids = new_or_updated_insertions
.iter()
.map(|i| i.transaction.id.clone().unwrap())
.collect::<Vec<_>>();
Ok((new_or_updated_insertions, inserted_transaction_ids))
} }
async fn notify_new_transactions( async fn notify_new_transactions(
@@ -114,16 +158,29 @@ async fn notify_new_transactions(
} }
mod tests { mod tests {
use super::{insert, notify_new_transactions, update_expenditures, update_transactions};
use crate::ingestion::ingestion_logic::from_json_row;
use anyhow::Error; use anyhow::Error;
use sea_orm::{DatabaseConnection}; use entity::account;
use sqlx::{PgPool};
use sqlx::postgres::PgListener;
use testcontainers::ContainerAsync;
use migration::MigratorTrait; use migration::MigratorTrait;
use sea_orm::{ActiveModelTrait, DatabaseConnection, TransactionTrait};
use serde_json::Value;
use sqlx::postgres::PgListener;
use sqlx::{Executor, PgPool};
use testcontainers::runners::AsyncRunner; use testcontainers::runners::AsyncRunner;
use super::notify_new_transactions; use testcontainers::ContainerAsync;
use tokio::sync::OnceCell;
async fn initialise() -> Result<(ContainerAsync<testcontainers_modules::postgres::Postgres>, DatabaseConnection, PgPool), Error> { #[derive(Debug)]
struct DatabaseInstance {
container: ContainerAsync<testcontainers_modules::postgres::Postgres>,
db: DatabaseConnection,
pool: PgPool,
}
static INSTANCE: OnceCell<DatabaseInstance> = OnceCell::const_new();
async fn initialise_db() -> Result<DatabaseInstance, Error> {
let container = testcontainers_modules::postgres::Postgres::default() let container = testcontainers_modules::postgres::Postgres::default()
.start() .start()
.await?; .await?;
@@ -137,30 +194,143 @@ mod tests {
let db: DatabaseConnection = sea_orm::Database::connect(connection_string).await?; let db: DatabaseConnection = sea_orm::Database::connect(connection_string).await?;
migration::Migrator::up(&db, None).await?; migration::Migrator::up(&db, None).await?;
let pool = PgPool::connect(connection_string) let pool = PgPool::connect(connection_string).await?;
.await?; let instance = DatabaseInstance {
container,
db,
pool,
};
Ok((container, db, pool)) Ok(instance)
}
async fn get_or_initialize_db_instance() -> Result<&'static DatabaseInstance, Error> {
Ok(INSTANCE
.get_or_init(|| async { initialise_db().await.unwrap() })
.await)
}
async fn create_test_account(db: &DatabaseConnection) -> Result<i32, Error> {
let new_account = account::ActiveModel {
id: sea_orm::ActiveValue::NotSet,
name: sea_orm::ActiveValue::Set("Test Account".to_string()),
};
let inserted = new_account.insert(db).await?;
Ok(inserted.id)
}
#[tokio::test]
async fn test_empty_insertion_list() -> Result<(), Error> {
let db = get_or_initialize_db_instance().await?;
let insertions = vec![];
let tx = db.db.begin().await?;
update_transactions(&tx, &insertions).await?;
update_expenditures(&tx, &insertions, &vec![]).await?;
tx.commit().await?;
Ok(())
} }
#[tokio::test] #[tokio::test]
async fn test_notify() -> Result<(), Error> { async fn test_notify() -> Result<(), Error> {
let (_container, db, pool) = initialise().await?; let dbi = get_or_initialize_db_instance().await?;
let mut listener = PgListener::connect_with(&pool).await?; let mut listener = PgListener::connect_with(&dbi.pool).await?;
listener.listen("monzo_new_transactions").await?; listener.listen("monzo_new_transactions").await?;
let ids = vec!["test1".to_string(), "test2".to_string(), "test3".to_string()]; let ids = vec![
"test1".to_string(),
"test2".to_string(),
"test3".to_string(),
];
notify_new_transactions( notify_new_transactions(&dbi.db, &ids).await?;
&db,
&ids,
).await?;
let notification = listener.recv().await?; let notification = listener.recv().await?;
let payload = notification.payload(); let payload = notification.payload();
println!("Payload: {}", payload); println!("Payload: {}", payload);
assert_eq!(serde_json::from_str::<Vec<String>>(&payload)?, ids, "Payloads do not match"); assert_eq!(
serde_json::from_str::<Vec<String>>(&payload)?,
ids,
"Payloads do not match"
);
Ok(())
}
#[tokio::test]
async fn test_notify_on_insert() -> Result<(), Error> {
let dbi = get_or_initialize_db_instance().await?;
let account_id = create_test_account(&dbi.db).await?;
let mut listener = PgListener::connect_with(&dbi.pool).await?;
listener.listen("monzo_new_transactions").await?;
let json = include_str!("../../fixtures/transactions.json");
let json: Vec<Vec<Value>> = serde_json::from_str(json).unwrap();
let data = json
.iter()
.map(|row| from_json_row(row))
.collect::<Result<Vec<_>, anyhow::Error>>()?;
insert(&dbi.db, data.clone(), account_id).await?;
let notification = listener.recv().await?;
let payload = notification.payload();
let mut payload = serde_json::from_str::<Vec<String>>(&payload)?;
payload.sort();
let mut ids = data
.iter()
.map(|row| row.transaction_id.clone())
.collect::<Vec<_>>();
ids.sort();
assert_eq!(payload, ids, "Inserted IDs do not match");
insert(&dbi.db, data.clone(), account_id).await?;
let notification = listener.recv().await?;
let payload = notification.payload();
let payload = serde_json::from_str::<Vec<String>>(&payload)?;
assert_eq!(
payload,
Vec::<String>::new(),
"Re-inserting identical rows triggered double notification"
);
let mut altered_data = data.clone();
altered_data[0].description = Some("New description".to_string());
assert_ne!(
altered_data[0].compute_hash(),
data[0].compute_hash(),
"Alterations have the same hash"
);
insert(&dbi.db, altered_data.clone(), account_id).await?;
let notification = listener.recv().await?;
let payload = notification.payload();
let payload = serde_json::from_str::<Vec<String>>(&payload)?;
assert_eq!(
payload,
vec![altered_data[0].transaction_id.clone()],
"Re-inserting altered row failed to re-trigger notification"
);
Ok(()) Ok(())
} }
} }
pub(crate) async fn get_account_id(
p0: &DatabaseConnection,
p1: Option<String>,
) -> Result<i32, AppError> {
let p1 = p1.unwrap_or("Monzo".to_string());
entity::prelude::Account::find()
.filter(entity::account::Column::Name.eq(p1))
.select_only()
.column(entity::account::Column::Id)
.into_tuple::<i32>()
.one(p0)
.await?
.ok_or(AppError::BadRequest(anyhow!("Account not found")))
}
+9
View File
@@ -0,0 +1,9 @@
#[allow(dead_code)]
pub mod headings {
#[allow(unused_imports)]
pub use super::super::ingestion_logic::headings::*;
// Additional Flex headings
pub const MONEY_OUT: usize = 16;
pub const MONEY_IN: usize = 17;
}
+148 -97
View File
@@ -1,16 +1,18 @@
use crate::ingestion::db::Insertion; use crate::ingestion::db::Insertion;
use anyhow::Context; use anyhow::{anyhow, Context};
use chrono::{DateTime, NaiveDate, NaiveDateTime, NaiveTime}; use chrono::{DateTime, NaiveDate, NaiveDateTime, NaiveTime};
use csv::StringRecord; use csv::StringRecord;
use entity::expenditure::ActiveModel; use entity::expenditure::ActiveModel;
use entity::transaction; use entity::transaction;
use num_traits::FromPrimitive; use num_traits::FromPrimitive;
use sea_orm::prelude::Decimal; use sea_orm::prelude::Decimal;
use sea_orm::ActiveValue::*;
use sea_orm::IntoActiveModel; use sea_orm::IntoActiveModel;
use serde_json::Value;
use std::hash::Hash;
use crate::ingestion::flex;
#[allow(dead_code)] #[allow(dead_code)]
mod headings { pub(crate) mod headings {
pub const TRANSACTION_ID: usize = 0; pub const TRANSACTION_ID: usize = 0;
pub const DATE: usize = 1; pub const DATE: usize = 1;
pub const TIME: usize = 2; pub const TIME: usize = 2;
@@ -29,7 +31,23 @@ mod headings {
pub const CATEGORY_SPLIT: usize = 15; pub const CATEGORY_SPLIT: usize = 15;
} }
fn parse_section(monzo_transaction_id: &str, section: &str) -> anyhow::Result<ActiveModel> { #[derive(Debug, Eq, PartialEq, Hash, Clone)]
pub struct MonzoRow {
pub category_split: Option<String>,
pub primary_category: String,
pub total_amount: Decimal,
pub receipt: Option<String>,
pub notes: Option<String>,
pub emoji: Option<String>,
pub description: Option<String>,
pub transaction_type: String,
pub title: Option<String>,
pub timestamp: NaiveDateTime,
pub transaction_id: String,
}
impl MonzoRow {
fn parse_section(monzo_transaction_id: &str, section: &str) -> anyhow::Result<ActiveModel> {
let mut components = section.split(':'); let mut components = section.split(':');
let category: String = components let category: String = components
.next() .next()
@@ -47,23 +65,82 @@ fn parse_section(monzo_transaction_id: &str, section: &str) -> anyhow::Result<Ac
amount, amount,
} }
.into_active_model()) .into_active_model())
}
/// Compute a hash of this row, returning the number as an i64 to be used as a unique constraint
/// in the database.
pub fn compute_hash(&self) -> i64 {
use std::collections::hash_map::DefaultHasher;
use std::hash::Hasher;
let mut hasher = DefaultHasher::new();
self.hash(&mut hasher);
hasher.finish() as i64
}
pub fn into_insertion(self, account_id: i32) -> Result<Insertion, anyhow::Error> {
let identity_hash = self.compute_hash();
let expenditures: Vec<_> = match &self.category_split {
Some(split) if !split.is_empty() => split
.split(',')
.map(|section| Self::parse_section(&self.transaction_id, section))
.collect::<Result<Vec<_>, anyhow::Error>>()?,
_ => vec![entity::expenditure::Model {
category: self.primary_category.clone(),
amount: self.total_amount,
transaction_id: self.transaction_id.clone(),
}
.into_active_model()],
};
Ok(Insertion {
transaction: transaction::Model {
id: self.transaction_id,
transaction_type: self.transaction_type,
timestamp: self.timestamp,
title: self.title,
emoji: self.emoji,
notes: self.notes,
receipt: self.receipt,
total_amount: self.total_amount,
description: self.description,
identity_hash: Some(identity_hash),
account_id: Some(account_id),
}
.into_active_model(),
contained_expenditures: expenditures,
identity_hash,
})
}
} }
fn json_opt(value: &serde_json::Value) -> Option<String> { fn json_opt(value: &Value) -> Option<String> {
match value { match value {
serde_json::Value::String(string) if string.is_empty() => None, Value::String(string) if string.is_empty() => None,
serde_json::Value::String(string) => Some(string.to_string()), Value::String(string) => Some(string.to_string()),
_ => None, _ => None,
} }
} }
pub fn from_json_row(row: Vec<serde_json::Value>) -> anyhow::Result<Insertion> { fn json_required_str(value: &Value, label: &str) -> anyhow::Result<String> {
use serde_json::Value; match value {
let monzo_transaction_id = row[headings::TRANSACTION_ID] Value::String(string) if string.is_empty() => Err(anyhow!("{} is empty", label)),
.as_str() Value::String(string) => Ok(string.to_string()),
.context("No transaction id")? _ => Err(anyhow!("{} is not a string", label)),
.to_string(); }
}
fn parse_timestamp(date: &str, time: &str) -> anyhow::Result<NaiveDateTime> {
let date = NaiveDate::parse_from_str(date, "%Y-%m-%d")?;
let time = NaiveTime::parse_from_str(time, "%H:%M:%S")?;
Ok(date.and_time(time))
}
pub fn from_json_row(row: &[Value]) -> anyhow::Result<MonzoRow> {
let date = DateTime::parse_from_rfc3339(row[headings::DATE].as_str().context("No date")?) let date = DateTime::parse_from_rfc3339(row[headings::DATE].as_str().context("No date")?)
.context("Failed to parse date")?; .context("Failed to parse date")?;
@@ -73,57 +150,57 @@ pub fn from_json_row(row: Vec<serde_json::Value>) -> anyhow::Result<Insertion> {
let timestamp = date.date_naive().and_time(time); let timestamp = date.date_naive().and_time(time);
let title = row[headings::NAME]
.as_str()
.context("No title")?
.to_string();
let monzo_transaction_type = row[headings::TYPE]
.as_str()
.context("No transaction type")?
.to_string();
let description = json_opt(&row[headings::DESCRIPTION]);
let emoji = json_opt(&row[headings::EMOJI]);
let notes = json_opt(&row[headings::NOTES_AND_TAGS]);
let receipt = json_opt(&row[headings::RECEIPT]);
let total_amount = Decimal::from_f64(row[headings::AMOUNT].as_f64().context("No amount")?) let total_amount = Decimal::from_f64(row[headings::AMOUNT].as_f64().context("No amount")?)
.context("Failed to parse date")?; .context("Failed to parse date")?;
let expenditures: Vec<_> = match row.get(headings::CATEGORY_SPLIT) { Ok(MonzoRow {
Some(Value::String(split)) if !split.is_empty() => split transaction_id: json_required_str(&row[headings::TRANSACTION_ID], "Transaction ID")?,
.split(',') title: json_opt(&row[headings::NAME]),
.map(|section| parse_section(&monzo_transaction_id, section)) transaction_type: json_required_str(&row[headings::TYPE], "Transaction type")?,
.collect::<Result<Vec<_>, anyhow::Error>>()?, description: json_opt(&row[headings::DESCRIPTION]),
emoji: json_opt(&row[headings::EMOJI]),
_ => vec![entity::expenditure::Model { notes: json_opt(&row[headings::NOTES_AND_TAGS]),
category: row[headings::CATEGORY] receipt: json_opt(&row[headings::RECEIPT]),
.as_str() primary_category: json_required_str(&row[headings::CATEGORY], "Primary Category")?,
.context("No context")? category_split: json_opt(&row[headings::CATEGORY_SPLIT]),
.to_string(), total_amount,
amount: total_amount, timestamp,
transaction_id: monzo_transaction_id.clone(),
}
.into_active_model()],
};
Ok(Insertion {
transaction: transaction::ActiveModel {
id: Set(monzo_transaction_id),
transaction_type: Set(monzo_transaction_type),
timestamp: Set(timestamp),
title: Set(title),
emoji: Set(emoji),
notes: Set(notes),
receipt: Set(receipt),
total_amount: Set(total_amount),
description: Set(description),
},
contained_expenditures: expenditures,
}) })
} }
#[test]
fn test_json() {
let json = include_str!("../../fixtures/transactions.json");
let csv = include_str!("../../fixtures/transactions.csv");
let json: Vec<Vec<Value>> = serde_json::from_str(json).unwrap();
let mut csv_reader = csv::Reader::from_reader(csv.as_bytes());
let json_rows = json
.iter()
.map(|row| from_json_row(&row))
.collect::<Result<Vec<_>, anyhow::Error>>()
.unwrap();
let csv_rows = csv_reader
.records()
.map(|record| from_csv_row(record.unwrap()))
.collect::<Result<Vec<_>, anyhow::Error>>()
.unwrap();
assert_eq!(csv_rows.len(), json_rows.len(), "Different number of rows");
for (i, (json_row, csv_row)) in json_rows.iter().zip(csv_rows.iter()).enumerate() {
assert_eq!(json_row, csv_row, "Row {} is different", i);
assert_eq!(
json_row.compute_hash(),
csv_row.compute_hash(),
"Row {} hash are different",
i
);
}
}
fn csv_opt(s: &str) -> Option<String> { fn csv_opt(s: &str) -> Option<String> {
match s { match s {
"" => None, "" => None,
@@ -131,9 +208,7 @@ fn csv_opt(s: &str) -> Option<String> {
} }
} }
pub fn from_csv_row(row: StringRecord) -> anyhow::Result<Insertion> { pub fn from_csv_row(row: StringRecord) -> anyhow::Result<MonzoRow> {
let monzo_transaction_id = row[headings::TRANSACTION_ID].to_string();
let date = NaiveDate::parse_from_str(&row[headings::DATE], "%d/%m/%Y") let date = NaiveDate::parse_from_str(&row[headings::DATE], "%d/%m/%Y")
.context("Failed to parse date from csv")?; .context("Failed to parse date from csv")?;
@@ -142,41 +217,17 @@ pub fn from_csv_row(row: StringRecord) -> anyhow::Result<Insertion> {
let timestamp = NaiveDateTime::new(date, time); let timestamp = NaiveDateTime::new(date, time);
let title = row[headings::NAME].to_string(); Ok(MonzoRow {
let monzo_transaction_type = row[headings::TYPE].to_string(); timestamp,
transaction_id: row[headings::TRANSACTION_ID].to_string(),
let description = csv_opt(&row[headings::DESCRIPTION]); title: csv_opt(&row[headings::NAME]),
let emoji = csv_opt(&row[headings::EMOJI]); transaction_type: row[headings::TYPE].to_string(),
let notes = csv_opt(&row[headings::NOTES_AND_TAGS]); description: csv_opt(&row[headings::DESCRIPTION]),
let receipt = csv_opt(&row[headings::RECEIPT]); emoji: csv_opt(&row[headings::EMOJI]),
let total_amount = row[headings::AMOUNT].parse::<Decimal>()?; notes: csv_opt(&row[headings::NOTES_AND_TAGS]),
receipt: csv_opt(&row[headings::RECEIPT]),
let expenditures: Vec<_> = match row.get(headings::CATEGORY_SPLIT) { total_amount: row[headings::AMOUNT].parse::<Decimal>()?,
Some(split) if !split.is_empty() => split category_split: csv_opt(&row[headings::CATEGORY_SPLIT]),
.split(',') primary_category: row[headings::CATEGORY].to_string(),
.map(|section| parse_section(&monzo_transaction_id, section))
.collect::<Result<Vec<_>, anyhow::Error>>()?,
_ => vec![entity::expenditure::Model {
transaction_id: monzo_transaction_id.clone(),
category: row[headings::CATEGORY].to_string(),
amount: total_amount,
}
.into_active_model()],
};
Ok(Insertion {
transaction: transaction::ActiveModel {
id: Set(monzo_transaction_id),
transaction_type: Set(monzo_transaction_type),
timestamp: Set(timestamp),
title: Set(title),
emoji: Set(emoji),
notes: Set(notes),
receipt: Set(receipt),
total_amount: Set(total_amount),
description: Set(description),
},
contained_expenditures: expenditures,
}) })
} }
+1
View File
@@ -1,3 +1,4 @@
pub mod db; pub mod db;
pub mod ingestion_logic; pub mod ingestion_logic;
pub mod flex;
pub mod routes; pub mod routes;
+98 -31
View File
@@ -2,66 +2,133 @@ use crate::error::AppError;
use crate::ingestion::db; use crate::ingestion::db;
use crate::ingestion::ingestion_logic::{from_csv_row, from_json_row}; use crate::ingestion::ingestion_logic::{from_csv_row, from_json_row};
use anyhow::anyhow; use anyhow::anyhow;
use axum::extract::multipart::MultipartError;
use axum::extract::{Extension, Json, Multipart}; use axum::extract::{Extension, Json, Multipart};
use bytes::Bytes;
use sea_orm::DatabaseConnection; use sea_orm::DatabaseConnection;
use serde_json::Value; use serde_json::Value;
use std::io::Cursor; use std::io::Cursor;
pub async fn monzo_updated( #[derive(serde::Deserialize, Debug)]
Extension(db): Extension<DatabaseConnection>, #[serde(untagged)]
Json(row): Json<Vec<Value>>, pub enum MonzoBatchedJsonInput {
) -> Result<&'static str, AppError> { Legacy(Vec<Vec<Value>>),
db::insert(&db, vec![from_json_row(row)?]).await.unwrap(); New {
account_id: Option<u8>,
rows: Vec<Vec<Value>>,
},
}
Ok("Ok") impl MonzoBatchedJsonInput {
fn account_id(&self) -> Option<u8> {
match self {
MonzoBatchedJsonInput::Legacy(_) => None,
MonzoBatchedJsonInput::New { account_id, .. } => *account_id,
}
}
fn rows(&self) -> &[Vec<Value>] {
match self {
MonzoBatchedJsonInput::Legacy(rows) => rows,
MonzoBatchedJsonInput::New { rows, .. } => rows,
}
}
} }
pub async fn monzo_batched_json( pub async fn monzo_batched_json(
Extension(db): Extension<DatabaseConnection>, Extension(db): Extension<DatabaseConnection>,
Json(data): Json<Vec<Vec<Value>>>, Json(data): Json<MonzoBatchedJsonInput>,
) -> Result<&'static str, AppError> { ) -> Result<&'static str, AppError> {
let insertions = data let rows = data
.into_iter() .rows()
.skip(1) .iter()
.map(|row| from_json_row(row)) .skip_while(|row| row[0] == Value::String("Transaction ID".to_string()))
.map(|row| from_json_row(row.as_ref()))
.collect::<Result<_, _>>()?; .collect::<Result<_, _>>()?;
db::insert(&db, insertions).await.unwrap(); // We default to the main account for JSON ingestion for now.
db::insert(&db, rows, data.account_id().unwrap_or(1) as i32).await?;
Ok("Ok")
}
async fn extract_csv_and_account_name(
mut multipart: Multipart,
) -> Result<(Option<Bytes>, Option<String>), MultipartError> {
let mut csv = None;
let mut account_name = None;
while let Some(field) = multipart.next_field().await? {
match field.name() {
Some("csv") => {
csv = Some(field.bytes().await?);
}
Some("account_id") => {
account_name = Some(field.text().await?);
}
_ => {}
}
if csv.is_some() && account_name.is_some() {
break;
}
}
Ok((csv, account_name))
}
#[derive(serde::Deserialize, Debug, Clone)]
pub struct ShortcutBody {
pub body: String,
pub account_name: String,
}
pub async fn shortcuts_csv(
Extension(db): Extension<DatabaseConnection>,
Json(shortcut_body): Json<ShortcutBody>,
) -> Result<&'static str, AppError> {
let account_id = db::get_account_id(&db, Some(shortcut_body.account_name)).await?;
let csv = Cursor::new(shortcut_body.body.as_bytes());
let mut csv = csv::Reader::from_reader(csv);
let data = csv.records();
let data = data
.filter_map(|f| f.ok())
.map(from_csv_row)
.collect::<Result<_, _>>()?;
db::insert(&db, data, account_id).await?;
Ok("Ok") Ok("Ok")
} }
pub async fn monzo_batched_csv( pub async fn monzo_batched_csv(
Extension(db): Extension<DatabaseConnection>, Extension(db): Extension<DatabaseConnection>,
mut multipart: Multipart, multipart: Multipart,
) -> Result<&'static str, AppError> { ) -> Result<&'static str, AppError> {
let csv = loop { static CSV_MISSING_ERR_MSG: &str = "No CSV file provided. Expected a multipart request with a `csv` field containing the contents of the CSV.";
match multipart.next_field().await.unwrap() {
Some(field) if field.name() == Some("csv") => {
break Some(field.bytes().await.unwrap());
}
Some(_) => {} let (csv, account_name) = extract_csv_and_account_name(multipart)
None => break None, .await
} .map_err(|e| AppError::BadRequest(anyhow!(e)))?;
};
let Some(csv) = csv else { let Some(csv) = csv else {
return Err(AppError::BadRequest(anyhow!("No CSV file provided"))); return Err(AppError::BadRequest(anyhow!(CSV_MISSING_ERR_MSG)));
}; };
let account_id = db::get_account_id(&db, account_name).await?;
let csv = Cursor::new(csv); let csv = Cursor::new(csv);
let mut csv = csv::Reader::from_reader(csv); let mut csv = csv::Reader::from_reader(csv);
let data = csv.records(); let data = csv.records();
let data = data
db::insert( .filter_map(|f| f.ok())
&db,
data.filter_map(|f| f.ok())
.map(from_csv_row) .map(from_csv_row)
.collect::<Result<_, _>>()?, .collect::<Result<_, _>>()?;
)
.await db::insert(&db, data, account_id).await?;
.unwrap();
Ok("Ok") Ok("Ok")
} }
+91 -18
View File
@@ -2,27 +2,64 @@ mod error;
mod ingestion; mod ingestion;
use crate::error::AppError; use crate::error::AppError;
use crate::ingestion::routes::{monzo_batched_csv, monzo_batched_json, monzo_updated}; use crate::ingestion::db;
use crate::ingestion::ingestion_logic::from_csv_row;
use crate::ingestion::routes::{monzo_batched_csv, monzo_batched_json, shortcuts_csv};
use axum::routing::{get, post}; use axum::routing::{get, post};
use axum::{Extension, Router}; use axum::{Extension, Router};
use clap::Parser; use clap::{Parser, Subcommand};
use migration::{Migrator, MigratorTrait}; use migration::{Migrator, MigratorTrait};
use sea_orm::{ConnectionTrait, DatabaseConnection}; use sea_orm::{ConnectionTrait, DatabaseConnection};
use std::fs::File;
use std::net::SocketAddr; use std::net::SocketAddr;
use std::path::PathBuf;
use tower_http::trace::TraceLayer;
use tracing::log::LevelFilter;
#[derive(Debug, clap::Parser)] #[derive(Debug, Subcommand)]
struct Config { enum Commands {
/// If we should perform migration on startup. /// Manually run database migrations.
Migrate {
/// Number of migration steps to perform. If not provided, all migrations will be run.
#[arg(long)]
steps: Option<u32>,
/// If we should perform migration down.
#[arg(long)]
down: bool,
},
/// Start web app for to the google-sheets app script
Serve {
/// If we should perform migration at startup.
#[clap(short, long, env, default_value_t = true)] #[clap(short, long, env, default_value_t = true)]
migrate: bool, migrate: bool,
/// The server address to bind to. /// The server address to bind to.
#[clap(short, long, env, default_value = "0.0.0.0:3000")] #[clap(short, long, env, default_value = "0.0.0.0:3000")]
addr: SocketAddr, addr: SocketAddr,
},
/// Ingest a google-sheets CSV export into the database.
Csv {
/// The path of the CSV file to ingest.
csv_file: PathBuf,
/// The name of the account to ingest the CSV for.
#[clap(long, short)]
account: String,
},
}
/// Ingest and manage monzo transactions from the Monzo Plus auto-export feature
#[derive(Debug, clap::Parser)]
struct Cli {
/// URL to PostgreSQL database. /// URL to PostgreSQL database.
#[clap(short, long = "db", env)] #[clap(short, long = "db", env)]
database_url: String, database_url: String,
#[command(subcommand)]
command: Commands,
} }
async fn health_check( async fn health_check(
@@ -35,24 +72,60 @@ async fn health_check(
#[tokio::main] #[tokio::main]
async fn main() -> anyhow::Result<()> { async fn main() -> anyhow::Result<()> {
let config: Config = Config::parse(); tracing_subscriber::fmt::init();
let connection = sea_orm::Database::connect(&config.database_url).await?;
if config.migrate { let cli: Cli = Cli::parse();
let connection = sea_orm::ConnectOptions::new(&cli.database_url)
.sqlx_logging_level(LevelFilter::Debug)
.to_owned();
let connection = sea_orm::Database::connect(connection).await?;
match cli.command {
Commands::Migrate { steps, down } => {
if down {
Migrator::down(&connection, steps).await?;
} else {
Migrator::up(&connection, steps).await?
}
}
Commands::Serve { migrate, addr } => {
if migrate {
Migrator::up(&connection, None).await?; Migrator::up(&connection, None).await?;
} }
tracing_subscriber::fmt::init(); serve_web(addr, connection).await?;
let app = Router::new() }
.route("/health", get(health_check))
.route("/monzo-updated", post(monzo_updated))
.route("/monzo-batch-export", post(monzo_batched_json))
.route("/monzo-csv-ingestion", post(monzo_batched_csv))
.layer(Extension(connection.clone()));
tracing::debug!("listening on {}", &config.addr); Commands::Csv { csv_file, account: account_name } => {
let listener = tokio::net::TcpListener::bind(&config.addr).await.unwrap(); let mut csv = csv::Reader::from_reader(File::open(csv_file)?);
axum::serve(listener, app).await.unwrap(); let data = csv.records();
let data = data
.filter_map(|f| f.ok())
.map(from_csv_row)
.collect::<Result<_, _>>()?;
let account_id = db::get_account_id(&connection, Some(account_name)).await?;
db::insert(&connection, data, account_id).await?;
}
}
Ok(())
}
async fn serve_web(address: SocketAddr, connection: DatabaseConnection) -> anyhow::Result<()> {
let app = Router::new()
.route("/health", get(health_check))
.route("/monzo-batch-export", post(monzo_batched_json))
.route("/monzo-csv-ingestion", post(monzo_batched_csv))
.route("/shortcuts-csv-import", post(shortcuts_csv))
.layer(Extension(connection.clone()))
.layer(TraceLayer::new_for_http());
tracing::info!("listening on {}", &address);
let listener = tokio::net::TcpListener::bind(&address).await?;
axum::serve(listener, app).await?;
Ok(()) Ok(())
} }