File size: 3,337 Bytes
0283e25
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
---
license: apache-2.0
---
# Fred Test Corpus

**Purpose:** Benchmark and regression testing for the **Fred** open-source agentic AI platform ([https://fredk8.dev](https://fredk8.dev)).

---

## Overview

This dataset contains a **heterogeneous corpus of PDF documents** used to validate and benchmark the ingestion, vectorization, and retrieval (RAG) components of the Fred platform.

It includes real-world documents from three distinct domains:

| Folder | Source | Description |
|:--------|:--------|:-------------|
| **`ECB/`** | European Central Bank | Recent ECB Working Papers (economic research). |
| **`OCDE/`** | Organisation for Economic Co-operation and Development | French-language policy and analysis papers. |
| **`ARXIV-AI/`** | arXiv.org (Artificial Intelligence category) | AI-related scientific preprints from September 2025. |

Each folder contains between 8 and 15 representative PDF files, with typical file sizes between **300 KB and 2 MB**.

---

## ⚙️ Intended Use

The corpus is designed for **functional and performance testing** of Fred’s document ingestion and retrieval subsystems, including:

- **Document ingestion pipeline:** testing metadata extraction, text conversion, and storage backends (e.g. MinIO, OpenSearch, DuckDB).  
- **Embedding & vectorization:** validating chunking, embedding quality, and retrieval accuracy.  
- **RAG agents:** verifying contextual retrieval within Fred’s *DocumentsExpert* and *Knowledge Flow* agents.  
- **Library selection & tagging:** evaluating how agents choose contextually relevant sources from mixed-domain libraries.

This corpus is *not* a production dataset; it is meant for open-source experimentation, demos, and reproducibility of Fred agent behavior.

---

## Folder Structure

```
fred-test-corpus/
├── ECB/
│   ├── ecb.wp3098~f8a4989211.en.pdf
│   ├── ecb.wp3099~d0ab147fbf.en.pdf
│   └── ...
├── OCDE/
│   ├── 04b537a4-fr.pdf
│   ├── 299d5974-fr.pdf
│   └── ...
└── ARXIV-AI/
    ├── 2509.03768v1.pdf
    ├── 2509.03827v1.pdf
    └── ...
```

---

## ⚖️ Licensing and Sources

All included PDFs are **publicly available documents** from:
- [European Central Bank](https://www.ecb.europa.eu/)
- [Organisation de Coopération et de Développement Économiques (OCDE)](https://www.oecd.org/)
- [arXiv.org](https://arxiv.org/)

They are redistributed **for non-commercial research and testing purposes only**, under fair-use / open-access terms of their respective publishers.

> Fred and its contributors make no claim of ownership over the contents of these documents.  
> This corpus is distributed *solely for reproducibility and testing* of open-source AI workflows.

---

## 🧾 Citation

If you use this corpus in your work, please cite:

```
Fred Open Source Project (2025).  
"Fred Test Corpus for Document Ingestion and RAG Evaluation."  
https://huggingface.co/fredk8/fred-test-corpus
```

---

## 🧱 Related Projects

- Fred Agentic FRamework](https://github.com/ThalesGroup/fred)
- [Official Site](https://fredk8.dev)

---

**Maintainer:** [ThalesGroup / Fred Open Source](https://github.com/ThalesGroup/fred)  
**License:** Apache 2.0 (for scripts & metadata)  
**Data License:** As per original publishers (non-commercial research use)