Medyassino commited on
Commit
42e1ab3
·
verified ·
1 Parent(s): b429d6c

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +162 -0
README.md ADDED
@@ -0,0 +1,162 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ task_categories:
4
+ - text-generation
5
+ - question-answering
6
+ language:
7
+ - fr
8
+ - en
9
+ tags:
10
+ - marketplace
11
+ - ecommerce
12
+ - products
13
+ - digital-marketplace
14
+ - french
15
+ - jsonl
16
+ pretty_name: "MarketplaceAI JSONL Dataset"
17
+ size_categories:
18
+ - 10K<n<100K
19
+ ---
20
+
21
+ # MarketplaceJSONLDataset - Dataset JSONL
22
+
23
+ Dataset au format JSON Lines contenant des données de produits pour l'entraînement de modèles d'IA spécialisés dans les marketplaces.
24
+
25
+ ## 📊 Statistiques du Dataset
26
+
27
+ - **Format** : JSON Lines (.jsonl)
28
+ - **Nombre total d'exemples** : 9,599,999
29
+ - **Taille totale** : 10912.11 MB
30
+ - **Nombre de fichiers** : 1
31
+
32
+ ## 📁 Structure des Fichiers
33
+
34
+ ### test.jsonl
35
+ - **Lignes** : 9,599,999
36
+ - **Taille** : 10912.11 MB
37
+ - **Champs** : product_id, category_id, seo_title_fr, name_fr, created_at, product_type, description_fr, tags_en, comments_en, tags_fr, description_en, seo_description_fr, variants, seo_description_en, regular_price, comments_fr, seo_title_en, updated_at, extend_price, discount_offer, name_en
38
+
39
+ **Exemple :**
40
+ ```json
41
+ {
42
+ "product_id": "59ddbb01-64d7-487d-bf9c-d2359572c8b7",
43
+ "name_en": "Le plaisir de concrétiser vos projets plus rapidement",
44
+ "name_fr": "Le confort d'avancer à sa source",
45
+ "description_en": "<p>Parvenir poitrine empêcher voir seulement pas je. High-quality digital solution.</p>",
46
+ "description_fr": "<p>Seulement vaincre chanter aussi servir lors rose. Solution numérique de haute qualité.</p>",
47
+ "tags_en": "digital, api, learning",
48
+ "tags_fr": "apprentissage, tutoriel, automatisation",
49
+ "seo_title_en": "Le plaisir de concrétiser vos projets plus rapidement - Best modèle de simulation 3D Resource",
50
+ "seo_title_fr": "Le confort d'avancer à sa source - Meilleure ressource template",
51
+ "seo_description_en": "Discover Le plaisir de concrétiser vos projets plus rapidement, the ultimate visualisation de données interactive solution.",
52
+ "seo_description_fr": "Découvrez Le confort d'avancer à sa source, la solution ultime educational_content.",
53
+ "regular_price": 140.67,
54
+ "extend_price": 184.06,
55
+ "category_id": 11,
56
+ "product_type": "module d'intégration API",
57
+ "comments_en": "Social nul. | Chiffre intéresser professeur.",
58
+ "comments_fr": "Port.",
59
+ "discount_offer": "10% OFF",
60
+ "variants": "Standard ($140.67) | Extended ($184.06)",
61
+ "created_at": "2025-04-03 20:35:27",
62
+ "updated_at": "2025-05-22 08:53:09"
63
+ }
64
+ ```
65
+
66
+
67
+ ## 🔧 Format des Données
68
+
69
+ Chaque ligne du fichier JSONL contient un objet JSON valide :
70
+
71
+ ```jsonl
72
+ {"id": 1, "name": "Produit A", "description": "Description du produit...", "price": 99.99}
73
+ {"id": 2, "name": "Produit B", "description": "Autre description...", "price": 149.99}
74
+ ```
75
+
76
+ ## 💻 Utilisation
77
+
78
+ ### Avec Python
79
+ ```python
80
+ import json
81
+
82
+ # Lire un fichier JSONL
83
+ def read_jsonl(file_path):
84
+ with open(file_path, 'r', encoding='utf-8') as f:
85
+ for line in f:
86
+ yield json.loads(line.strip())
87
+
88
+ # Utilisation
89
+ for item in read_jsonl('produits_numeriquesv2.jsonl'):
90
+ print(item)
91
+ ```
92
+
93
+ ### Avec Pandas
94
+ ```python
95
+ import pandas as pd
96
+ import json
97
+
98
+ # Charger un JSONL en DataFrame
99
+ def jsonl_to_dataframe(file_path):
100
+ data = []
101
+ with open(file_path, 'r', encoding='utf-8') as f:
102
+ for line in f:
103
+ data.append(json.loads(line.strip()))
104
+ return pd.DataFrame(data)
105
+
106
+ df = jsonl_to_dataframe('produits_numeriquesv2.jsonl')
107
+ print(df.head())
108
+ ```
109
+
110
+ ### Avec MarketplaceAI
111
+ ```python
112
+ from marketplace_ai_server import MultilingualMarketplaceBot
113
+
114
+ # Convertir JSONL en JSON pour l'entraînement
115
+ def convert_jsonl_to_json(jsonl_path, json_path):
116
+ products = []
117
+ with open(jsonl_path, 'r', encoding='utf-8') as f:
118
+ for line in f:
119
+ products.append(json.loads(line.strip()))
120
+
121
+ with open(json_path, 'w', encoding='utf-8') as f:
122
+ json.dump({"products": products}, f, indent=2, ensure_ascii=False)
123
+
124
+ # Convertir et entraîner
125
+ convert_jsonl_to_json('produits_numeriquesv2.jsonl', 'products.json')
126
+
127
+ bot = MultilingualMarketplaceBot()
128
+ bot.train_on_json_file('products.json')
129
+ ```
130
+
131
+ ## 🚀 Intégration avec Hugging Face Datasets
132
+
133
+ ```python
134
+ from datasets import load_dataset
135
+
136
+ # Charger le dataset
137
+ dataset = load_dataset('Medyassino/MarketplaceJSONLDataset')
138
+
139
+ # Accéder aux données
140
+ for example in dataset['train']:
141
+ print(example)
142
+ ```
143
+
144
+ ## 📚 Citation
145
+
146
+ Si vous utilisez ce dataset, veuillez citer :
147
+
148
+ ```bibtex
149
+ @dataset{marketplace_jsonl_dataset,
150
+ title={MarketplaceAI JSONL Dataset},
151
+ author={Medyassino},
152
+ year={2025},
153
+ format={JSONL},
154
+ url={https://huggingface.co/datasets/Medyassino/MarketplaceJSONLDataset}
155
+ }
156
+ ```
157
+
158
+ ## 🔗 Ressources Connexes
159
+
160
+ - [Modèle MarketplaceAI](https://huggingface.co/Medyassino/YassinoQuantum)
161
+ - [Dataset JSON](https://huggingface.co/datasets/Medyassino/MarketplaceDataset)
162
+ - [Documentation du format JSONL](http://jsonlines.org/)