Merge remote-tracking branch 'origin/main' into robin/thesis/experime…

…nt-configs

.gitignore

@@ -1,3 +1,7 @@
+# Images
+*.svg
+*.png
+
 # Logging files
 *.log
 

.pylintrc

@@ -186,8 +186,9 @@ disable=raw-checker-failed,
         too-many-arguments,             # we can't determine a good limit here. reviews should spot bad cases of this.
         duplicate-code,                 # Mostly imports and test setup.
         cyclic-import,                  # We use these inside methods that require models from multiple apps. Tests will catch actual errors.
-        too-many-instance-attributes    # We always ignore this anyways
-
+        too-many-instance-attributes,   # We always ignore this anyways
+        too-many-positional-arguments,  # We do not want to limit the number of positional arguments
+        too-many-locals                 # We always ignore this anyways
 # Enable the message, report, category or checker with the given id(s). You can
 # either give multiple identifier separated by comma (,) or put this option
 # multiple time (only on the command line, not in the configuration file where

README.md

@@ -7,7 +7,7 @@
 [![codecov](https://codecov.io/github/eth-easl/modyn/graph/badge.svg?token=KFDCE03SQ4)](https://codecov.io/github/eth-easl/modyn)
 [![License](https://img.shields.io/github/license/eth-easl/modyn)](https://img.shields.io/github/license/eth-easl/modyn)
 
-Modyn is an open-source platform for model training on growing datasets, i.e., datasets where points get added over time.
+Modyn is a data-centric machine learning pipeline orchestrator, i.e., a platform for model training on growing datasets where points get added over time. Check out our [blog post](https://systems.ethz.ch/research/blog/modyn.html) for a brief introduction.
 
 </div>
 
@@ -55,9 +55,8 @@ For running all integration tests, run
 Checkout our [Example Pipeline](docs/EXAMPLE.md) guide for an example on how to run a Modyn pipeline.
 Checkout our [Technical Guidelines](docs/TECHNICAL.md) for some hints on developing Modyn and how to add new data selection and triggering policies.
 Checkout the [Architecture Documentation](docs/ARCHITECTURE.md) for an overview of Modyn's components.
-Last, checkout our [vision paper on Modyn](https://anakli.inf.ethz.ch/papers/MLonDynamicData_EuroMLSys23.pdf) for an introduction to model training on dynamic datasets.
+Last, checkout our [full paper on Modyn](https://anakli.inf.ethz.ch/papers/modyn_sigmod25.pdf) for more technical background and experiments we ran using Modyn.
 
-We are actively developing and designing Modyn, including more thorough documentation.
 Please reach out via Github, Twitter, E-Mail, or any other channel of communication if you are interested in collaborating, have any questions, or have any problems running Modyn.
 
 How to [contribute](docs/CONTRIBUTING.md).
@@ -81,3 +80,16 @@ We welcome input from both research and practice.
 
 Modyn is being developed at the [Efficient Architectures and Systems Lab (EASL)](https://anakli.inf.ethz.ch/#Group) at the [ETH Zurich Systems Group](https://systems.ethz.ch/).
 Please reach out to `mboether [at] inf [­dot] ethz [dot] ch` or open an issue on Github if you have any questions or inquiry related to Modyn and its usage.
+
+### Paper / Citation
+
+If you use Modyn, please cite our SIGMOD'25 paper:
+
+```bibtex
+@inproceedings{Bother2025Modyn,
+  author = {B\"{o}ther, Maximilian and Robroek, Ties and Gsteiger, Viktor and Ma, Xianzhe and T\"{o}z\"{u}n, P{\i}nar and Klimovic, Ana},
+  title = {Modyn: Data-Centric Machine Learning Pipeline Orchestration},
+  booktitle = {Proceedings of the Conference on Management of Data (SIGMOD)},
+  year = {2025},
+}
+```

_typos.toml

@@ -5,3 +5,4 @@ extend-ignore-re = ["(?Rm)^.*# spellchecker:disable-line$"]
 [default.extend-words]
 strat = "strat"
 fpr = "fpr"
+ther = "ther"

analytics/app/pages/plots/cost_vs_eval_metric_agg.py

@@ -1,5 +1,5 @@
 from dataclasses import dataclass
-from typing import get_args
+from typing import Any, get_args
 
 import pandas as pd
 import plotly.express as px
@@ -32,9 +32,9 @@ class _PageState:
 
 def gen_fig_scatter_num_triggers(
     page: str,
-    eval_handler: str,
-    dataset_id: str,
-    metric: str,
+    eval_handler: str | Any | None,
+    dataset_id: str | Any | None,
+    metric: str | Any | None,
     agg_func_x: AGGREGATION_FUNCTION,
     agg_func_y: EVAL_AGGREGATION_FUNCTION,
     stages: list[str],

analytics/app/pages/plots/eval_heatmap.py

@@ -1,4 +1,5 @@
 from dataclasses import dataclass
+from typing import Any
 
 import pandas as pd
 from dash import Input, Output, callback, dcc, html
@@ -31,9 +32,9 @@ def gen_figure(
     page: str,
     multi_pipeline_mode: bool,
     patch_yearbook: bool,
-    eval_handler: str,
-    dataset_id: str,
-    metric: str,
+    eval_handler: str | Any | None,
+    dataset_id: str | Any | None,
+    metric: str | Any | None,
 ) -> go.Figure:
     """Create the cost over time figure with barplot or histogram. Histogram
     has nice binning while barplot is precise.

analytics/app/pages/plots/eval_over_time.py

@@ -1,4 +1,5 @@
 from dataclasses import dataclass
+from typing import Any
 
 import pandas as pd
 import plotly.express as px
@@ -29,9 +30,9 @@ def gen_figure(
     page: str,
     multi_pipeline_mode: bool,
     patch_yearbook: bool,
-    eval_handler: str,
-    dataset_id: str,
-    metric: str,
+    eval_handler: str | Any | None,
+    dataset_id: str | Any | None,
+    metric: str | Any | None,
 ) -> go.Figure:
     """Create the evaluation over time figure with a line plot.
 

analytics/app/pages/plots/num_samples.py

@@ -34,12 +34,12 @@ class _PageState:
 def gen_figure(
     page: str,
     multi_pipeline_mode: bool,
-    time_metric: str,
-    y_axis: YAxis,
-    use_scatter_size: bool,
+    time_metric: str | None,
+    y_axis: YAxis | None,
+    use_scatter_size: bool | None,
     patch_yearbook: bool,
-    dataset_id: str,
-    eval_handler: str,
+    dataset_id: str | None,
+    eval_handler: str | None,
 ) -> go.Figure:
     """Create the cost over time figure with barplot or histogram. Histogram
     has nice binning while barplot is precise.

analytics/app/pages/plots/num_triggers_eval_metric.py

@@ -31,9 +31,9 @@ class _PageState:
 def gen_fig_scatter_num_triggers(
     page: str,
     multi_pipeline_mode: bool,
-    eval_handler: str,
-    dataset_id: str,
-    metric: str,
+    eval_handler: str | None,
+    dataset_id: str | None,
+    metric: str | None,
     aggregate_metric: bool = True,
     time_weighted: bool = True,
     only_active_periods: bool = True,