Spaces:
Sleeping
Sleeping
Ilyas KHIAT
commited on
Commit
•
e346593
1
Parent(s):
bdd1430
test
Browse files- .chainlit/config.toml +97 -0
- .chainlit/translations/en-US.json +155 -0
- .chainlit/translations/pt-BR.json +155 -0
- Dockerfile +26 -0
- chainlit.md +10 -0
- rag_app.py +69 -0
- rag_module.py +221 -0
- requirements.txt +9 -0
.chainlit/config.toml
ADDED
@@ -0,0 +1,97 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
[project]
|
2 |
+
# Whether to enable telemetry (default: true). No personal data is collected.
|
3 |
+
enable_telemetry = true
|
4 |
+
|
5 |
+
|
6 |
+
# List of environment variables to be provided by each user to use the app.
|
7 |
+
user_env = []
|
8 |
+
|
9 |
+
# Duration (in seconds) during which the session is saved when the connection is lost
|
10 |
+
session_timeout = 3600
|
11 |
+
|
12 |
+
# Enable third parties caching (e.g LangChain cache)
|
13 |
+
cache = false
|
14 |
+
|
15 |
+
# Authorized origins
|
16 |
+
allow_origins = ["*"]
|
17 |
+
|
18 |
+
# Follow symlink for asset mount (see https://github.com/Chainlit/chainlit/issues/317)
|
19 |
+
# follow_symlink = false
|
20 |
+
|
21 |
+
[features]
|
22 |
+
# Show the prompt playground
|
23 |
+
prompt_playground = true
|
24 |
+
|
25 |
+
# Process and display HTML in messages. This can be a security risk (see https://stackoverflow.com/questions/19603097/why-is-it-dangerous-to-render-user-generated-html-or-javascript)
|
26 |
+
unsafe_allow_html = false
|
27 |
+
|
28 |
+
# Process and display mathematical expressions. This can clash with "$" characters in messages.
|
29 |
+
latex = false
|
30 |
+
|
31 |
+
# Authorize users to upload files with messages
|
32 |
+
multi_modal = true
|
33 |
+
|
34 |
+
# Allows user to use speech to text
|
35 |
+
[features.speech_to_text]
|
36 |
+
enabled = false
|
37 |
+
# See all languages here https://github.com/JamesBrill/react-speech-recognition/blob/HEAD/docs/API.md#language-string
|
38 |
+
# language = "en-US"
|
39 |
+
|
40 |
+
[UI]
|
41 |
+
# Name of the app and chatbot.
|
42 |
+
name = "Chatbot"
|
43 |
+
|
44 |
+
# Show the readme while the thread is empty.
|
45 |
+
show_readme_as_default = true
|
46 |
+
|
47 |
+
# Description of the app and chatbot. This is used for HTML tags.
|
48 |
+
# description = ""
|
49 |
+
|
50 |
+
# Large size content are by default collapsed for a cleaner ui
|
51 |
+
default_collapse_content = true
|
52 |
+
|
53 |
+
# The default value for the expand messages settings.
|
54 |
+
default_expand_messages = false
|
55 |
+
|
56 |
+
# Hide the chain of thought details from the user in the UI.
|
57 |
+
hide_cot = false
|
58 |
+
|
59 |
+
# Link to your github repo. This will add a github button in the UI's header.
|
60 |
+
# github = ""
|
61 |
+
|
62 |
+
# Specify a CSS file that can be used to customize the user interface.
|
63 |
+
# The CSS file can be served from the public directory or via an external link.
|
64 |
+
# custom_css = "/public/test.css"
|
65 |
+
|
66 |
+
# Specify a Javascript file that can be used to customize the user interface.
|
67 |
+
# The Javascript file can be served from the public directory.
|
68 |
+
# custom_js = "/public/test.js"
|
69 |
+
|
70 |
+
# Specify a custom font url.
|
71 |
+
# custom_font = "https://fonts.googleapis.com/css2?family=Inter:wght@400;500;700&display=swap"
|
72 |
+
|
73 |
+
# Override default MUI light theme. (Check theme.ts)
|
74 |
+
[UI.theme]
|
75 |
+
#font_family = "Inter, sans-serif"
|
76 |
+
[UI.theme.light]
|
77 |
+
#background = "#FAFAFA"
|
78 |
+
#paper = "#FFFFFF"
|
79 |
+
|
80 |
+
[UI.theme.light.primary]
|
81 |
+
#main = "#F80061"
|
82 |
+
#dark = "#980039"
|
83 |
+
#light = "#FFE7EB"
|
84 |
+
|
85 |
+
# Override default MUI dark theme. (Check theme.ts)
|
86 |
+
[UI.theme.dark]
|
87 |
+
#background = "#FAFAFA"
|
88 |
+
#paper = "#FFFFFF"
|
89 |
+
|
90 |
+
[UI.theme.dark.primary]
|
91 |
+
#main = "#F80061"
|
92 |
+
#dark = "#980039"
|
93 |
+
#light = "#FFE7EB"
|
94 |
+
|
95 |
+
|
96 |
+
[meta]
|
97 |
+
generated_by = "1.0.301"
|
.chainlit/translations/en-US.json
ADDED
@@ -0,0 +1,155 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"components": {
|
3 |
+
"atoms": {
|
4 |
+
"buttons": {
|
5 |
+
"userButton": {
|
6 |
+
"menu": {
|
7 |
+
"settings": "Settings",
|
8 |
+
"settingsKey": "S",
|
9 |
+
"APIKeys": "API Keys",
|
10 |
+
"logout": "Logout"
|
11 |
+
}
|
12 |
+
}
|
13 |
+
}
|
14 |
+
},
|
15 |
+
"molecules": {
|
16 |
+
"newChatButton": {
|
17 |
+
"newChat": "New Chat"
|
18 |
+
},
|
19 |
+
"tasklist": {
|
20 |
+
"TaskList": {
|
21 |
+
"title": "\ud83d\uddd2\ufe0f Task List",
|
22 |
+
"loading": "Loading...",
|
23 |
+
"error": "An error occured"
|
24 |
+
}
|
25 |
+
},
|
26 |
+
"attachments": {
|
27 |
+
"cancelUpload": "Cancel upload",
|
28 |
+
"removeAttachment": "Remove attachment"
|
29 |
+
},
|
30 |
+
"newChatDialog": {
|
31 |
+
"createNewChat": "Create new chat?",
|
32 |
+
"clearChat": "This will clear the current messages and start a new chat.",
|
33 |
+
"cancel": "Cancel",
|
34 |
+
"confirm": "Confirm"
|
35 |
+
},
|
36 |
+
"settingsModal": {
|
37 |
+
"expandMessages": "Expand Messages",
|
38 |
+
"hideChainOfThought": "Hide Chain of Thought",
|
39 |
+
"darkMode": "Dark Mode"
|
40 |
+
}
|
41 |
+
},
|
42 |
+
"organisms": {
|
43 |
+
"chat": {
|
44 |
+
"history": {
|
45 |
+
"index": {
|
46 |
+
"lastInputs": "Last Inputs",
|
47 |
+
"noInputs": "Such empty...",
|
48 |
+
"loading": "Loading..."
|
49 |
+
}
|
50 |
+
},
|
51 |
+
"inputBox": {
|
52 |
+
"input": {
|
53 |
+
"placeholder": "Type your message here..."
|
54 |
+
},
|
55 |
+
"speechButton": {
|
56 |
+
"start": "Start recording",
|
57 |
+
"stop": "Stop recording"
|
58 |
+
},
|
59 |
+
"SubmitButton": {
|
60 |
+
"sendMessage": "Send message",
|
61 |
+
"stopTask": "Stop Task"
|
62 |
+
},
|
63 |
+
"UploadButton": {
|
64 |
+
"attachFiles": "Attach files"
|
65 |
+
},
|
66 |
+
"waterMark": {
|
67 |
+
"text": "Built with"
|
68 |
+
}
|
69 |
+
},
|
70 |
+
"Messages": {
|
71 |
+
"index": {
|
72 |
+
"running": "Running",
|
73 |
+
"executedSuccessfully": "executed successfully",
|
74 |
+
"failed": "failed",
|
75 |
+
"feedbackUpdated": "Feedback updated",
|
76 |
+
"updating": "Updating"
|
77 |
+
}
|
78 |
+
},
|
79 |
+
"dropScreen": {
|
80 |
+
"dropYourFilesHere": "Drop your files here"
|
81 |
+
},
|
82 |
+
"index": {
|
83 |
+
"failedToUpload": "Failed to upload",
|
84 |
+
"cancelledUploadOf": "Cancelled upload of",
|
85 |
+
"couldNotReachServer": "Could not reach the server",
|
86 |
+
"continuingChat": "Continuing previous chat"
|
87 |
+
},
|
88 |
+
"settings": {
|
89 |
+
"settingsPanel": "Settings panel",
|
90 |
+
"reset": "Reset",
|
91 |
+
"cancel": "Cancel",
|
92 |
+
"confirm": "Confirm"
|
93 |
+
}
|
94 |
+
},
|
95 |
+
"threadHistory": {
|
96 |
+
"sidebar": {
|
97 |
+
"filters": {
|
98 |
+
"FeedbackSelect": {
|
99 |
+
"feedbackAll": "Feedback: All",
|
100 |
+
"feedbackPositive": "Feedback: Positive",
|
101 |
+
"feedbackNegative": "Feedback: Negative"
|
102 |
+
},
|
103 |
+
"SearchBar": {
|
104 |
+
"search": "Search"
|
105 |
+
}
|
106 |
+
},
|
107 |
+
"DeleteThreadButton": {
|
108 |
+
"confirmMessage": "This will delete the thread as well as it's messages and elements.",
|
109 |
+
"cancel": "Cancel",
|
110 |
+
"confirm": "Confirm",
|
111 |
+
"deletingChat": "Deleting chat",
|
112 |
+
"chatDeleted": "Chat deleted"
|
113 |
+
},
|
114 |
+
"index": {
|
115 |
+
"pastChats": "Past Chats"
|
116 |
+
},
|
117 |
+
"ThreadList": {
|
118 |
+
"empty": "Empty..."
|
119 |
+
},
|
120 |
+
"TriggerButton": {
|
121 |
+
"closeSidebar": "Close sidebar",
|
122 |
+
"openSidebar": "Open sidebar"
|
123 |
+
}
|
124 |
+
},
|
125 |
+
"Thread": {
|
126 |
+
"backToChat": "Go back to chat",
|
127 |
+
"chatCreatedOn": "This chat was created on"
|
128 |
+
}
|
129 |
+
},
|
130 |
+
"header": {
|
131 |
+
"chat": "Chat",
|
132 |
+
"readme": "Readme"
|
133 |
+
}
|
134 |
+
}
|
135 |
+
},
|
136 |
+
"hooks": {
|
137 |
+
"useLLMProviders": {
|
138 |
+
"failedToFetchProviders": "Failed to fetch providers:"
|
139 |
+
}
|
140 |
+
},
|
141 |
+
"pages": {
|
142 |
+
"Design": {},
|
143 |
+
"Env": {
|
144 |
+
"savedSuccessfully": "Saved successfully",
|
145 |
+
"requiredApiKeys": "Required API Keys",
|
146 |
+
"requiredApiKeysInfo": "To use this app, the following API keys are required. The keys are stored on your device's local storage."
|
147 |
+
},
|
148 |
+
"Page": {
|
149 |
+
"notPartOfProject": "You are not part of this project."
|
150 |
+
},
|
151 |
+
"ResumeButton": {
|
152 |
+
"resumeChat": "Resume Chat"
|
153 |
+
}
|
154 |
+
}
|
155 |
+
}
|
.chainlit/translations/pt-BR.json
ADDED
@@ -0,0 +1,155 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"components": {
|
3 |
+
"atoms": {
|
4 |
+
"buttons": {
|
5 |
+
"userButton": {
|
6 |
+
"menu": {
|
7 |
+
"settings": "Configura\u00e7\u00f5es",
|
8 |
+
"settingsKey": "S",
|
9 |
+
"APIKeys": "Chaves de API",
|
10 |
+
"logout": "Sair"
|
11 |
+
}
|
12 |
+
}
|
13 |
+
}
|
14 |
+
},
|
15 |
+
"molecules": {
|
16 |
+
"newChatButton": {
|
17 |
+
"newChat": "Nova Conversa"
|
18 |
+
},
|
19 |
+
"tasklist": {
|
20 |
+
"TaskList": {
|
21 |
+
"title": "\ud83d\uddd2\ufe0f Lista de Tarefas",
|
22 |
+
"loading": "Carregando...",
|
23 |
+
"error": "Ocorreu um erro"
|
24 |
+
}
|
25 |
+
},
|
26 |
+
"attachments": {
|
27 |
+
"cancelUpload": "Cancelar envio",
|
28 |
+
"removeAttachment": "Remover anexo"
|
29 |
+
},
|
30 |
+
"newChatDialog": {
|
31 |
+
"createNewChat": "Criar novo chat?",
|
32 |
+
"clearChat": "Isso limpar\u00e1 as mensagens atuais e iniciar\u00e1 uma nova conversa.",
|
33 |
+
"cancel": "Cancelar",
|
34 |
+
"confirm": "Confirmar"
|
35 |
+
},
|
36 |
+
"settingsModal": {
|
37 |
+
"expandMessages": "Expandir Mensagens",
|
38 |
+
"hideChainOfThought": "Esconder Sequ\u00eancia de Pensamento",
|
39 |
+
"darkMode": "Modo Escuro"
|
40 |
+
}
|
41 |
+
},
|
42 |
+
"organisms": {
|
43 |
+
"chat": {
|
44 |
+
"history": {
|
45 |
+
"index": {
|
46 |
+
"lastInputs": "\u00daltimas Entradas",
|
47 |
+
"noInputs": "Vazio...",
|
48 |
+
"loading": "Carregando..."
|
49 |
+
}
|
50 |
+
},
|
51 |
+
"inputBox": {
|
52 |
+
"input": {
|
53 |
+
"placeholder": "Digite sua mensagem aqui..."
|
54 |
+
},
|
55 |
+
"speechButton": {
|
56 |
+
"start": "Iniciar grava\u00e7\u00e3o",
|
57 |
+
"stop": "Parar grava\u00e7\u00e3o"
|
58 |
+
},
|
59 |
+
"SubmitButton": {
|
60 |
+
"sendMessage": "Enviar mensagem",
|
61 |
+
"stopTask": "Parar Tarefa"
|
62 |
+
},
|
63 |
+
"UploadButton": {
|
64 |
+
"attachFiles": "Anexar arquivos"
|
65 |
+
},
|
66 |
+
"waterMark": {
|
67 |
+
"text": "Constru\u00eddo com"
|
68 |
+
}
|
69 |
+
},
|
70 |
+
"Messages": {
|
71 |
+
"index": {
|
72 |
+
"running": "Executando",
|
73 |
+
"executedSuccessfully": "executado com sucesso",
|
74 |
+
"failed": "falhou",
|
75 |
+
"feedbackUpdated": "Feedback atualizado",
|
76 |
+
"updating": "Atualizando"
|
77 |
+
}
|
78 |
+
},
|
79 |
+
"dropScreen": {
|
80 |
+
"dropYourFilesHere": "Solte seus arquivos aqui"
|
81 |
+
},
|
82 |
+
"index": {
|
83 |
+
"failedToUpload": "Falha ao enviar",
|
84 |
+
"cancelledUploadOf": "Envio cancelado de",
|
85 |
+
"couldNotReachServer": "N\u00e3o foi poss\u00edvel conectar ao servidor",
|
86 |
+
"continuingChat": "Continuando o chat anterior"
|
87 |
+
},
|
88 |
+
"settings": {
|
89 |
+
"settingsPanel": "Painel de Configura\u00e7\u00f5es",
|
90 |
+
"reset": "Redefinir",
|
91 |
+
"cancel": "Cancelar",
|
92 |
+
"confirm": "Confirmar"
|
93 |
+
}
|
94 |
+
},
|
95 |
+
"threadHistory": {
|
96 |
+
"sidebar": {
|
97 |
+
"filters": {
|
98 |
+
"FeedbackSelect": {
|
99 |
+
"feedbackAll": "Feedback: Todos",
|
100 |
+
"feedbackPositive": "Feedback: Positivo",
|
101 |
+
"feedbackNegative": "Feedback: Negativo"
|
102 |
+
},
|
103 |
+
"SearchBar": {
|
104 |
+
"search": "Buscar"
|
105 |
+
}
|
106 |
+
},
|
107 |
+
"DeleteThreadButton": {
|
108 |
+
"confirmMessage": "Isso deletar\u00e1 a conversa, assim como suas mensagens e elementos.",
|
109 |
+
"cancel": "Cancelar",
|
110 |
+
"confirm": "Confirmar",
|
111 |
+
"deletingChat": "Deletando conversa",
|
112 |
+
"chatDeleted": "Conversa deletada"
|
113 |
+
},
|
114 |
+
"index": {
|
115 |
+
"pastChats": "Conversas Anteriores"
|
116 |
+
},
|
117 |
+
"ThreadList": {
|
118 |
+
"empty": "Vazio..."
|
119 |
+
},
|
120 |
+
"TriggerButton": {
|
121 |
+
"closeSidebar": "Fechar barra lateral",
|
122 |
+
"openSidebar": "Abrir barra lateral"
|
123 |
+
}
|
124 |
+
},
|
125 |
+
"Thread": {
|
126 |
+
"backToChat": "Voltar para a conversa",
|
127 |
+
"chatCreatedOn": "Esta conversa foi criada em"
|
128 |
+
}
|
129 |
+
},
|
130 |
+
"header": {
|
131 |
+
"chat": "Conversa",
|
132 |
+
"readme": "Leia-me"
|
133 |
+
}
|
134 |
+
},
|
135 |
+
"hooks": {
|
136 |
+
"useLLMProviders": {
|
137 |
+
"failedToFetchProviders": "Falha ao buscar provedores:"
|
138 |
+
}
|
139 |
+
},
|
140 |
+
"pages": {
|
141 |
+
"Design": {},
|
142 |
+
"Env": {
|
143 |
+
"savedSuccessfully": "Salvo com sucesso",
|
144 |
+
"requiredApiKeys": "Chaves de API necess\u00e1rias",
|
145 |
+
"requiredApiKeysInfo": "Para usar este aplicativo, as seguintes chaves de API s\u00e3o necess\u00e1rias. As chaves s\u00e3o armazenadas localmente em seu dispositivo."
|
146 |
+
},
|
147 |
+
"Page": {
|
148 |
+
"notPartOfProject": "Voc\u00ea n\u00e3o faz parte deste projeto."
|
149 |
+
},
|
150 |
+
"ResumeButton": {
|
151 |
+
"resumeChat": "Continuar Conversa"
|
152 |
+
}
|
153 |
+
}
|
154 |
+
}
|
155 |
+
}
|
Dockerfile
ADDED
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
FROM python:3.10
|
2 |
+
|
3 |
+
WORKDIR /code
|
4 |
+
|
5 |
+
COPY ./requirements.txt /code/requirements.txt
|
6 |
+
|
7 |
+
RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
|
8 |
+
#RUN apt update && apt install -y ffmpeg
|
9 |
+
|
10 |
+
RUN useradd -m -u 1000 user
|
11 |
+
USER user
|
12 |
+
ENV HOME=/home/user \
|
13 |
+
PATH=/home/user/.local/bin:$PATH
|
14 |
+
|
15 |
+
WORKDIR $HOME/app
|
16 |
+
|
17 |
+
COPY --chown=user . $HOME/app
|
18 |
+
|
19 |
+
|
20 |
+
|
21 |
+
#COPY . .
|
22 |
+
#COPY .chainlit .chainlit
|
23 |
+
|
24 |
+
|
25 |
+
CMD ["chainlit", "run", "rag_app.py", "--host", "0.0.0.0", "--port", "7860"]
|
26 |
+
# CMD ["ls", "-a"]
|
chainlit.md
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Bienvenue sur le chatbot ASSISTANT PAC 🚀🤖
|
2 |
+
|
3 |
+
Bonjour et bienvenue,
|
4 |
+
|
5 |
+
Je suis un agent intelligent pour vous aider à trouver les aides financières proposées par la PAC.
|
6 |
+
Mes connaissances se basent sur la documentation officielle fournie par le Ministère de l'Agriculture et de la Souveraineté alimentaire sur la PAC 2023-2027.
|
7 |
+
Vous retrouverez ainsi des informations concernant les aides découplées et les aides couplées.
|
8 |
+
|
9 |
+
Posez votre question en fonction de votre statut et de votre spécialité.
|
10 |
+
|
rag_app.py
ADDED
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import re
|
2 |
+
import json
|
3 |
+
import chainlit as cl
|
4 |
+
|
5 |
+
from langchain_community.vectorstores import FAISS
|
6 |
+
from rag_module import RagModule
|
7 |
+
|
8 |
+
from collections import defaultdict
|
9 |
+
|
10 |
+
|
11 |
+
|
12 |
+
prompt_template = """
|
13 |
+
Tu t'appelles ASSISTANT PAC, agent intelligent spécialisé sur les aides financières agricoles, et ta mission est d'aider les agriculteurs (rices) et porteurs de projets agricoles à identifier les aides agricoles PAC disponibles.
|
14 |
+
Tu comprends et génère les réponses en français, jamais en anglais.
|
15 |
+
|
16 |
+
Merci de bien vouloir répondre aux questions en utilisant seulement le contexte suivant.
|
17 |
+
contexte: {context}
|
18 |
+
|
19 |
+
historique: {history}
|
20 |
+
|
21 |
+
question: {question}
|
22 |
+
réponse:
|
23 |
+
"""
|
24 |
+
|
25 |
+
|
26 |
+
|
27 |
+
##------------ CHAINLIT ---------------##
|
28 |
+
@cl.on_chat_start
|
29 |
+
async def start():
|
30 |
+
rag = RagModule()
|
31 |
+
|
32 |
+
db = rag.get_faiss_db()
|
33 |
+
|
34 |
+
qa_chain = rag.retrieval_qa_memory_chain(db, prompt_template)
|
35 |
+
|
36 |
+
msg = cl.Message(content="Lancement du bot...", author = "Assistant PAC")
|
37 |
+
await msg.send()
|
38 |
+
msg.content = "Bonjour et bienvenue sur le Chatbot spécialisé dans les aides de la PAC (Politique agricole commune). Posez directement votre question pour être conseillé ?"
|
39 |
+
await msg.update()
|
40 |
+
cl.user_session.set("chain", qa_chain)
|
41 |
+
|
42 |
+
@cl.on_message
|
43 |
+
async def main(message):
|
44 |
+
rag = RagModule()
|
45 |
+
|
46 |
+
chain = cl.user_session.get("chain")
|
47 |
+
|
48 |
+
cb = cl.AsyncLangchainCallbackHandler(
|
49 |
+
stream_final_answer = True,
|
50 |
+
answer_prefix_tokens=["FINAL", "ANSWER"]
|
51 |
+
)
|
52 |
+
|
53 |
+
cb.answer_reached=True
|
54 |
+
response = await chain.ainvoke(message.content, callbacks=[cb])
|
55 |
+
|
56 |
+
answer = response.get('result')
|
57 |
+
sources = rag.get_sources_document(response.get('source_documents'))
|
58 |
+
|
59 |
+
elements = [cl.Pdf(name = "Pdf", display ="inline", path = path) for path in sources]
|
60 |
+
|
61 |
+
|
62 |
+
if response.get('source_documents'):
|
63 |
+
answer = rag.shape_answer_with_source(answer, sources)
|
64 |
+
else:
|
65 |
+
answer += f"\nNo sources found"
|
66 |
+
|
67 |
+
await cl.Message(content=answer, elements=elements, author="Assistant PAC").send()
|
68 |
+
|
69 |
+
|
rag_module.py
ADDED
@@ -0,0 +1,221 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
|
2 |
+
#load & split data
|
3 |
+
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
4 |
+
# embed data
|
5 |
+
from langchain_mistralai import MistralAIEmbeddings
|
6 |
+
# vector store
|
7 |
+
from langchain_community.vectorstores import FAISS
|
8 |
+
# prompt
|
9 |
+
from langchain.prompts import PromptTemplate
|
10 |
+
# memory
|
11 |
+
from langchain.memory import ConversationBufferMemory
|
12 |
+
#llm
|
13 |
+
from langchain_mistralai.chat_models import ChatMistralAI
|
14 |
+
|
15 |
+
#chain modules
|
16 |
+
from langchain.chains import RetrievalQA
|
17 |
+
from langchain.embeddings import CacheBackedEmbeddings
|
18 |
+
from langchain.storage import LocalFileStore
|
19 |
+
|
20 |
+
from langchain_community.document_loaders import PyPDFLoader
|
21 |
+
|
22 |
+
# import PyPDF2
|
23 |
+
import os
|
24 |
+
import re
|
25 |
+
from dotenv import load_dotenv
|
26 |
+
load_dotenv()
|
27 |
+
from collections import defaultdict
|
28 |
+
|
29 |
+
api_key = os.environ.get("MISTRAL_API_KEY")
|
30 |
+
|
31 |
+
def extract_pdfs_from_folder(folder_path):
|
32 |
+
pdf_files = []
|
33 |
+
for file_name in os.listdir(folder_path):
|
34 |
+
if file_name.endswith(".pdf"):
|
35 |
+
pdf_files.append(os.path.join(folder_path, file_name))
|
36 |
+
|
37 |
+
extracted_texts = []
|
38 |
+
for pdf_file in pdf_files:
|
39 |
+
loader = PyPDFLoader(pdf_file)
|
40 |
+
pages = loader.load()
|
41 |
+
extracted_texts += pages
|
42 |
+
|
43 |
+
return extracted_texts
|
44 |
+
|
45 |
+
class RagModule():
|
46 |
+
def __init__(self):
|
47 |
+
self.mistral_api_key = api_key
|
48 |
+
self.model_name_embedding = "mistral-embed"
|
49 |
+
print(f"API KEY:, {self.mistral_api_key}")
|
50 |
+
self.embedding_model = MistralAIEmbeddings(model=self.model_name_embedding, mistral_api_key=self.mistral_api_key)
|
51 |
+
|
52 |
+
self.chunk_size = 1000
|
53 |
+
self.chunk_overlap = 120
|
54 |
+
self.text_splitter = RecursiveCharacterTextSplitter(chunk_size=self.chunk_size, chunk_overlap=self.chunk_overlap)
|
55 |
+
self.db_faiss_path = "data/vector_store"
|
56 |
+
#params llm
|
57 |
+
self.llm_model = "mistral-small"
|
58 |
+
self.max_new_tokens = 512
|
59 |
+
self.top_p = 0.5
|
60 |
+
self.temperature = 0.1
|
61 |
+
|
62 |
+
|
63 |
+
|
64 |
+
|
65 |
+
def split_text(self, text:str) -> list:
|
66 |
+
"""Split the text into chunk
|
67 |
+
|
68 |
+
Args:
|
69 |
+
text (str): _description_
|
70 |
+
|
71 |
+
Returns:
|
72 |
+
list: _description_
|
73 |
+
"""
|
74 |
+
texts = self.text_splitter.split_text(text)
|
75 |
+
return texts
|
76 |
+
|
77 |
+
def get_metadata(self, texts:list) -> list:
|
78 |
+
"""_summary_
|
79 |
+
|
80 |
+
Args:
|
81 |
+
texts (list): _description_
|
82 |
+
|
83 |
+
Returns:
|
84 |
+
list: _description_
|
85 |
+
"""
|
86 |
+
metadatas = [{"source": f'Paragraphe: {i}'} for i in range(len(texts))]
|
87 |
+
return metadatas
|
88 |
+
|
89 |
+
def get_faiss_db(self):
|
90 |
+
"""load local faiss vector store containing all embeddings
|
91 |
+
|
92 |
+
"""
|
93 |
+
data = extract_pdfs_from_folder("./data/")
|
94 |
+
|
95 |
+
text_splitter = RecursiveCharacterTextSplitter(
|
96 |
+
chunk_size=1000,
|
97 |
+
chunk_overlap=100
|
98 |
+
)
|
99 |
+
|
100 |
+
chunked_documents = text_splitter.split_documents(data)
|
101 |
+
embedding_model = MistralAIEmbeddings(model=self.model_name_embedding, mistral_api_key=self.mistral_api_key)
|
102 |
+
|
103 |
+
|
104 |
+
store = LocalFileStore("./cache/")
|
105 |
+
embedder = CacheBackedEmbeddings.from_bytes_store(embedding_model, store, namespace=embedding_model.model)
|
106 |
+
|
107 |
+
vector_store = FAISS.from_documents(chunked_documents, embedder)
|
108 |
+
vector_store.save_local("faiss_index")
|
109 |
+
|
110 |
+
return vector_store
|
111 |
+
|
112 |
+
def set_custom_prompt(self, prompt_template:str):
|
113 |
+
"""Instantiate prompt template for Q&A retreival for each vectore stores
|
114 |
+
|
115 |
+
Args:
|
116 |
+
prompt_template (str): description of the prompt
|
117 |
+
input_variables (list): variables in the prompt
|
118 |
+
"""
|
119 |
+
prompt = PromptTemplate.from_template(
|
120 |
+
template=prompt_template,
|
121 |
+
)
|
122 |
+
|
123 |
+
return prompt
|
124 |
+
|
125 |
+
def load_mistral(self):
|
126 |
+
"""instantiate LLM
|
127 |
+
"""
|
128 |
+
|
129 |
+
model_kwargs = {
|
130 |
+
"mistral_api_key": self.mistral_api_key,
|
131 |
+
"model": self.llm_model,
|
132 |
+
"max_new_tokens": self.max_new_tokens,
|
133 |
+
"top_p": self.top_p,
|
134 |
+
"temperature": self.temperature,
|
135 |
+
}
|
136 |
+
|
137 |
+
llm = ChatMistralAI(**model_kwargs)
|
138 |
+
|
139 |
+
return llm
|
140 |
+
|
141 |
+
def retrieval_qa_memory_chain(self, db, prompt_template):
|
142 |
+
"""_summary_
|
143 |
+
"""
|
144 |
+
llm = self.load_mistral()
|
145 |
+
prompt = self.set_custom_prompt(prompt_template)
|
146 |
+
memory = ConversationBufferMemory(
|
147 |
+
memory_key = 'history',
|
148 |
+
input_key = 'question'
|
149 |
+
)
|
150 |
+
chain_type_kwargs= {
|
151 |
+
"prompt" : prompt,
|
152 |
+
"memory" : memory
|
153 |
+
}
|
154 |
+
|
155 |
+
qa_chain = RetrievalQA.from_chain_type(
|
156 |
+
llm = llm,
|
157 |
+
chain_type = 'stuff',
|
158 |
+
retriever = db.as_retriever(search_kwargs={"k":5}),
|
159 |
+
chain_type_kwargs = chain_type_kwargs,
|
160 |
+
return_source_documents = True,
|
161 |
+
)
|
162 |
+
|
163 |
+
return qa_chain
|
164 |
+
|
165 |
+
def retrieval_qa_chain(self, db, prompt_template):
|
166 |
+
"""_summary_
|
167 |
+
"""
|
168 |
+
llm = self.load_llm()
|
169 |
+
prompt = self.set_custom_prompt(prompt_template)
|
170 |
+
|
171 |
+
chain_type_kwargs= {
|
172 |
+
"prompt" : prompt,
|
173 |
+
}
|
174 |
+
|
175 |
+
qa_chain = RetrievalQA.from_chain_type(
|
176 |
+
llm = llm,
|
177 |
+
chain_type = 'stuff',
|
178 |
+
retriever = db.as_retriever(search_kwargs={"k":3}),
|
179 |
+
chain_type_kwargs = chain_type_kwargs,
|
180 |
+
return_source_documents = True,
|
181 |
+
)
|
182 |
+
|
183 |
+
return qa_chain
|
184 |
+
|
185 |
+
|
186 |
+
|
187 |
+
def get_sources_document(self, source_documents:list) -> dict:
|
188 |
+
"""generate dictionnary with path (as a key) and list of pages associated to one path
|
189 |
+
|
190 |
+
Args:
|
191 |
+
source_document (list): list of documents containing source_document of rag response
|
192 |
+
|
193 |
+
Returns:
|
194 |
+
dict: {
|
195 |
+
path/to/file1 : [0, 1, 3],
|
196 |
+
path/to/file2 : [5, 2]
|
197 |
+
}
|
198 |
+
"""
|
199 |
+
sources = defaultdict(list)
|
200 |
+
for doc in source_documents:
|
201 |
+
sources[doc.metadata["source"]].append(doc.metadata["page"])
|
202 |
+
|
203 |
+
return sources
|
204 |
+
|
205 |
+
def shape_answer_with_source(self, answer: str, sources: dict):
|
206 |
+
"""_summary_
|
207 |
+
|
208 |
+
Args:
|
209 |
+
answer (str): _description_
|
210 |
+
source (dict): _description_
|
211 |
+
"""
|
212 |
+
pattern = r"^(.+)\/([^\/]+)$"
|
213 |
+
|
214 |
+
source_msg = ""
|
215 |
+
for path, page in sources.items():
|
216 |
+
file = re.findall(pattern, path)[0][1]
|
217 |
+
source_msg += f"\nFichier: {file} - Page: {page}"
|
218 |
+
|
219 |
+
answer += f"\n{source_msg}"
|
220 |
+
|
221 |
+
return answer
|
requirements.txt
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
langchain==0.1.7
|
2 |
+
langchain-core==0.1.23
|
3 |
+
langchain-mistralai==0.0.4
|
4 |
+
langchain-community==0.0.20
|
5 |
+
faiss-cpu==1.7.4
|
6 |
+
python-dotenv==1.0.1
|
7 |
+
chainlit
|
8 |
+
openai
|
9 |
+
pypdf==4.0.2
|