Skip to content

Commit f5ca27c

Browse files
souradipp76souradipp76
committed
Created server and added FAISS.
1 parent 03abaa6 commit f5ca27c

File tree

7 files changed

+266
-141
lines changed

7 files changed

+266
-141
lines changed

‎app/__init__.py‎

Whitespace-only changes.

‎app/app.py‎

Lines changed: 54 additions & 135 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,10 @@
11
import streamlit as st
22
import os
33
import shutil
4-
import git
54
from streamlit_ace import st_ace
6-
from pymarkdown.api import PyMarkdownApi
5+
import requests
6+
7+
from streamlit.runtime.scriptrunner import get_script_run_ctx
78

89
# App title
910
st.set_page_config(page_title="Readme Generator", layout="wide",page_icon=":material/graphic_eq:")
@@ -20,28 +21,18 @@
2021
hf_token = st.text_input('Enter HuggingFace token:', type='password')
2122
os.environ['HF_TOKEN'] = hf_token
2223

23-
from readme_ready.query import query
24-
from readme_ready.index import index
25-
from readme_ready.types import AutodocRepoConfig, AutodocUserConfig, LLMModels
24+
ctx = get_script_run_ctx()
25+
session_id = None
26+
if ctx:
27+
session_id = ctx.session_id
28+
# You can now use session_id for various purposes, e.g., logging or file naming
29+
print(f"Current Streamlit session ID: {session_id}")
30+
else:
31+
print("Not running within a Streamlit context.")
2632

2733
with st.form("my_form"):
2834
st.subheader('Model')
29-
options = [
30-
LLMModels.TINYLLAMA_1p1B_CHAT_GGUF.value,
31-
LLMModels.GOOGLE_GEMMA_2B_INSTRUCT_GGUF.value,
32-
LLMModels.LLAMA2_7B_CHAT_GPTQ.value,
33-
LLMModels.LLAMA2_13B_CHAT_GPTQ.value,
34-
LLMModels.CODELLAMA_7B_INSTRUCT_GPTQ.value,
35-
LLMModels.CODELLAMA_13B_INSTRUCT_GPTQ.value,
36-
LLMModels.LLAMA2_7B_CHAT_HF.value,
37-
LLMModels.LLAMA2_13B_CHAT_HF.value,
38-
LLMModels.CODELLAMA_7B_INSTRUCT_HF.value,
39-
LLMModels.CODELLAMA_13B_INSTRUCT_HF.value,
40-
LLMModels.GOOGLE_GEMMA_2B_INSTRUCT.value,
41-
LLMModels.GOOGLE_GEMMA_7B_INSTRUCT.value,
42-
LLMModels.GOOGLE_CODEGEMMA_2B.value,
43-
LLMModels.GOOGLE_CODEGEMMA_7B_INSTRUCT.value,
44-
]
35+
options = requests.get(url="http://127.0.0.1:8000/models").json()
4536
llm = st.selectbox('Choose a model', options, key='llm')
4637
device = st.selectbox('Choose a device', ["cpu", "auto"], key='device')
4738
st.subheader('Parameters')
@@ -59,115 +50,23 @@
5950
submitted = st.form_submit_button("Setup")
6051
if submitted:
6152
st.toast('Indexing repository...')
62-
try:
63-
repo = git.Repo.clone_from(project_url, project_root)
64-
except:
65-
print('Project already exists.')
66-
67-
match llm:
68-
case LLMModels.TINYLLAMA_1p1B_CHAT_GGUF.value:
69-
model = LLMModels.TINYLLAMA_1p1B_CHAT_GGUF
70-
case LLMModels.GOOGLE_GEMMA_2B_INSTRUCT_GGUF.value:
71-
model = LLMModels.GOOGLE_GEMMA_2B_INSTRUCT_GGUF
72-
case LLMModels.LLAMA2_7B_CHAT_GPTQ.value:
73-
model = LLMModels.LLAMA2_7B_CHAT_GPTQ
74-
case LLMModels.LLAMA2_13B_CHAT_GPTQ.value:
75-
model = LLMModels.LLAMA2_13B_CHAT_GPTQ
76-
case LLMModels.CODELLAMA_7B_INSTRUCT_GPTQ.value:
77-
model = LLMModels.CODELLAMA_7B_INSTRUCT_GPTQ
78-
case LLMModels.CODELLAMA_13B_INSTRUCT_GPTQ.value:
79-
model = LLMModels.CODELLAMA_13B_INSTRUCT_GPTQ
80-
case LLMModels.LLAMA2_13B_CHAT_HF.value:
81-
model = LLMModels.LLAMA2_13B_CHAT_HF
82-
case LLMModels.CODELLAMA_7B_INSTRUCT_HF.value:
83-
model = LLMModels.CODELLAMA_7B_INSTRUCT_HF
84-
case LLMModels.CODELLAMA_13B_INSTRUCT_HF.value:
85-
model = LLMModels.CODELLAMA_13B_INSTRUCT_HF
86-
case LLMModels.GOOGLE_GEMMA_2B_INSTRUCT.value:
87-
model = LLMModels.GOOGLE_GEMMA_2B_INSTRUCT
88-
case LLMModels.GOOGLE_GEMMA_7B_INSTRUCT.value:
89-
model = LLMModels.GOOGLE_GEMMA_7B_INSTRUCT
90-
case LLMModels.GOOGLE_CODEGEMMA_2B.value:
91-
model = LLMModels.GOOGLE_CODEGEMMA_2B
92-
case LLMModels.GOOGLE_CODEGEMMA_7B_INSTRUCT.value:
93-
model = LLMModels.GOOGLE_CODEGEMMA_7B_INSTRUCT
94-
case _:
95-
model = LLMModels.LLAMA2_7B_CHAT_HF
96-
97-
repo_config = {
53+
request_body = {
9854
"name": name,
99-
"root": project_root,
100-
"repository_url": project_url,
101-
"output": output_dir,
102-
"llms": [model],
103-
"peft_model_path": None,
104-
"ignore": [
105-
".*",
106-
"*package-lock.json",
107-
"*package.json",
108-
"node_modules",
109-
"*dist*",
110-
"*build*",
111-
"*test*",
112-
"*.svg",
113-
"*.md",
114-
"*.mdx",
115-
"*.toml"
116-
],
117-
"file_prompt": "Write a detailed technical explanation of \
118-
what this code does. \n Focus on the high-level \
119-
purpose of the code and how it may be used in the \
120-
larger project.\n Include code examples where \
121-
appropriate. Keep you response between 100 and 300 \
122-
words. \n DO NOT RETURN MORE THAN 300 WORDS.\n \
123-
Output should be in markdown format.\n \
124-
Do not just list the methods and classes in this file.",
125-
"folder_prompt": "Write a technical explanation of what the \
126-
code in this file does\n and how it might fit into the \
127-
larger project or work with other parts of the project.\n \
128-
Give examples of how this code might be used. Include code \
129-
examples where appropriate.\n Be concise. Include any \
130-
information that may be relevant to a developer who is \
131-
curious about this code.\n Keep you response under \
132-
400 words. Output should be in markdown format.\n \
133-
Do not just list the files and folders in this folder.",
134-
"chat_prompt": "",
135-
"content_type": "docs",
136-
"target_audience": "smart developer",
137-
"link_hosted": True,
138-
"priority": None,
139-
"max_concurrent_calls": 50,
140-
"add_questions": False,
55+
"project_url": project_url,
56+
"model": llm,
14157
"device": device,
142-
}
143-
user_config = {
144-
"llms": [model]
58+
"temperature": temperature,
59+
"top_p": top_p,
60+
"max_length": max_length,
61+
# "peft_model_path": None,
14562
}
14663

147-
repo_conf = AutodocRepoConfig(
148-
name=repo_config["name"],
149-
repository_url=repo_config["repository_url"],
150-
root=repo_config["root"],
151-
output=repo_config["output"],
152-
llms=repo_config["llms"],
153-
peft_model_path=repo_config["peft_model_path"],
154-
priority=repo_config["priority"],
155-
max_concurrent_calls=repo_config["max_concurrent_calls"],
156-
add_questions=repo_config["add_questions"],
157-
ignore=repo_config["ignore"],
158-
file_prompt=repo_config["file_prompt"],
159-
folder_prompt=repo_config["folder_prompt"],
160-
chat_prompt=repo_config["chat_prompt"],
161-
content_type=repo_config["content_type"],
162-
target_audience=repo_config["target_audience"],
163-
link_hosted=repo_config["link_hosted"],
164-
device=repo_config["device"]
165-
)
166-
usr_conf = AutodocUserConfig(llms=user_config['llms'], streaming=True)
167-
index.index(repo_conf)
168-
st.session_state.repo_conf = repo_conf
169-
st.session_state.usr_conf = usr_conf
170-
st.session_state.chain = query.init_readme_chain(st.session_state.repo_conf, st.session_state.usr_conf)
64+
response = requests.post(
65+
url="http://localhost:8000/setup",
66+
json=request_body,
67+
headers={"x-session-id": session_id}
68+
).json()
69+
17170
st.toast('Repository indexing done.')
17271

17372

@@ -200,16 +99,32 @@ def clear_chat_history():
20099
if st.session_state.messages[-1]["role"] != "assistant":
201100
with st.spinner("Thinking..."):
202101
with history.chat_message("assistant"):
203-
if "chain" not in st.session_state.keys():
204-
full_response = 'Please setup model and repository!!'
205-
else:
206-
chain = st.session_state.chain
102+
response = requests.post(
103+
url="http://localhost:8000/query",
104+
json={"query": prompt},
105+
headers={"x-session-id": session_id},
106+
stream=True
107+
)
108+
109+
if response.status_code == 200:
207110
placeholder = st.empty()
208111
full_response = ''
209-
for chunk in chain.stream({'input': prompt}):
210-
if answer_chunk := chunk.get("answer"):
211-
full_response += answer_chunk
212-
placeholder.markdown(full_response)
112+
for chunk in response.iter_content(chunk_size=10, decode_unicode=True):
113+
full_response += chunk
114+
placeholder.markdown(full_response)
115+
else:
116+
full_response = "Please setup model and repository!!"
117+
118+
# if "chain" not in st.session_state.keys():
119+
# full_response = 'Please setup model and repository!!'
120+
# else:
121+
# chain = st.session_state.chain
122+
# placeholder = st.empty()
123+
# full_response = ''
124+
# for chunk in chain.stream({'input': prompt}):
125+
# if answer_chunk := chunk.get("answer"):
126+
# full_response += answer_chunk
127+
# placeholder.markdown(full_response)
213128

214129
message = {"role": "assistant", "content": full_response}
215130
st.session_state.messages.append(message)
@@ -240,7 +155,11 @@ def clear_chat_history():
240155

241156
def validate_markdown():
242157
error_str = ""
243-
errors = PyMarkdownApi().scan_string(st.session_state.readme_content)
158+
# errors = PyMarkdownApi().scan_string(st.session_state.readme_content)
159+
errors = requests.get(
160+
url="http://localhost:8000/validate_markdown",
161+
json={"content": st.session_state.readme_content}
162+
).json()
244163
if len(errors.scan_failures) > 0:
245164
print(errors.scan_failures)
246165
error_str = "\n".join([f'Line {failure.line_number}: Col {failure.column_number}: {failure.rule_id}: {failure.rule_description} {failure.extra_error_information} ({failure.rule_name})' for failure in errors.scan_failures])

0 commit comments

Comments
 (0)