You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+27-12Lines changed: 27 additions & 12 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -9,7 +9,9 @@
9
9
10
10
**fluidize-python** is a library for building modular, reproducible scientific computing pipelines. It provides a unified interface to a wide range of physical simulation tools, eliminating the need to navigate the inconsistent, incomplete instructions that often vary from tool to tool.
11
11
12
-
This library marks our first step toward AI-orchestrated scientific computing. By standardizing tools and practices within our framework, AI agents can automatically build, configure, and execute computational pipelines across domains and simulation platforms. Our goal is to improve today’s simulation tools so AI can assist researchers and scientists in accelerating the pace of innovation and scientific discovery.
12
+
This library marks our first step toward AI-orchestrated scientific computing. By standardizing tools and practices within our framework, AI agents can automatically build, configure, and execute computational pipelines across domains and simulation platforms.
13
+
14
+
Our goal is to improve today’s simulation tools so AI can assist researchers and scientists in accelerating the pace of innovation and scientific discovery.
13
15
14
16
## Quick Start
15
17
@@ -54,6 +56,7 @@ Students and researchers face significant barriers when working with different s
54
56
-**Reproducibility issues** – Sharing and reproducing experiments is frequently cumbersome and error-prone.
55
57
-**Scaling friction** – Moving from a local prototype to a cloud environment or dedicated compute cluster can be slow and difficult.
56
58
59
+
57
60
## The Solution
58
61
59
62
Fluidize provides a standardized wrapper that turns complex scientific software into modular components. This makes it possible to:
@@ -64,10 +67,16 @@ Fluidize provides a standardized wrapper that turns complex scientific software
64
67
65
68
All of this works with **minimal or no changes** to the existing codebase, allowing our framework to scale effortlessly to any repository.
66
69
70
+
67
71
## Architecture
68
72
73
+
At Fluidize, we believe strong organization leads to better reproducibility and scalability.
74
+
75
+
We treat each simulation pipeline as an individual project. Within projects, each pipeline is treated as a DAG (directed acyclic graph), where nodes represent individual pieces of scientific software (e.g. inputs, solvers, visualization tools, etc.) and edges represent data flow between nodes.
76
+
77
+
69
78
### Nodes
70
-
The foundational building blocks of Fluidize. Each node encapsulates a computational unit with:
79
+
Nodes are the foundational building blocks of simulation pipelines. Each node represents a computational unit with:
71
80
72
81
| File | Purpose |
73
82
|------|---------|
@@ -84,18 +93,20 @@ The foundational building blocks of Fluidize. Each node encapsulates a computati
84
93
- No source code modification required
85
94
- Automated node generation support (Public launch soon)
86
95
96
+
87
97
### Projects
88
-
The project currently hosts a simple layer for composing and managing multiple nodes:
98
+
99
+
Projects store a simple data layer for managing individual modules within a pipeline.
89
100
90
101
| File | Purpose |
91
102
|------|---------|
92
-
|`graph.json`| Node connectivity and data flow definition|
103
+
|`graph.json`| Node (scientific software) and edge (data flow) definitions|
93
104
|`metadata.yaml`| Project description and configuration |
94
105
95
106
96
-
Docker engine is used for local execution. With API calls, we use the Kubernetes engine with Argo Workflow Manager.
97
-
107
+
### Runs
98
108
109
+
Pipelines can be executed both locally and on the cloud. Local execution is handled by Docker engine. Cloud execution is routed through our API, and uses the Kubernetes engine with Argo Workflow Manager.
99
110
100
111
101
112
## Documentation
@@ -107,18 +118,22 @@ Comprehensive documentation is available at [https://Fluidize-Inc.github.io/flui
We would love contributions and collaborations! Please see our [Contributing Guide](CONTRIBUTING.md) for details.
124
+
We would love to collaborate with you! Please see our [Contributing Guide](CONTRIBUTING.md) for details.
125
+
126
+
Also - we would love to help streamline your pipeline! Please reach out to us at [founders@fluidize.ai](mailto:founders@fluidize.ai).
127
+
113
128
114
-
Also - we would love to help streamline your research pipeline! Please reach out at [henry@fluidize.ai](mailto:henry@fluidize.ai) or [henrybae@g.harvard.edu](mailto:henrybae@g.harvard.edu).
129
+
## Vision and Roadmap
115
130
116
-
## Roadmap
131
+
This is just the beginning of what we believe will be a really exciting new era for how we conduct research and make discoveries in science.
117
132
118
-
This is just the beginning of what we think is a really exciting new era for how we learn science and do research. We will be releasing the following tools built from this framework:
133
+
By standardizing tools, we hope to significantly increase the effectiveness of AI in research and discovery. Soon, we will be releasing the following tools built from this framework:
119
134
120
-
-**Fluidize Playground**: Automatically explore and build simulation pipelines with natural language.
121
-
-**Auto-Fluidize**: Automatically convert obscure scientific software to run anywhere
135
+
-**Auto-Fluidize**: Automatically convert any scientific software to run anywhere with our framework.
136
+
-**Fluidize AI Playground**: Explore and build simulation pipelines with natural language.
The Fluidize library is composed of a set of core modules that provide a high-level interface for managing Fluidize resources. These modules are designed to be used together to build and execute scientific computing pipelines.
4
+
5
+
## [Client](client.md)
6
+
7
+
The **Fluidize Client** provides a unified, high-level interface for managing Fluidize resources in both local and cloud API modes. It serves as the primary entry point for creating and running pipelines across these environments.
8
+
9
+
## [Projects](projects.md)
10
+
11
+
The **Projects** module provides tools for managing project lifecycles:
Focuses on individual projects, managing the project graph, nodes, and runs, and supporting execution of project-specific workflows.
18
+
19
+
## [Graph](graph.md)
20
+
21
+
The **Graph** module provides tools for managing the project graph, which is a representation of the simulation pipeline.
22
+
23
+
In a Fluidize project, pipelines are represented as a directed acyclic graph (DAG) where each node represents a module simulation and each edge represents the flow of data between nodes:
0 commit comments