Init repo
This commit is contained in:
parent
d39268cf49
commit
4b52d4aa91
|
|
@ -0,0 +1,10 @@
|
|||
env/*
|
||||
databases/storage.db
|
||||
4b751a4425c2884286a92fde2de6427f_trusted.table
|
||||
4b751a4425c2884286a92fde2de6427f_analitic.table
|
||||
4b751a4425c2884286a92fde2de6427f_analitic_voice.table
|
||||
4b751a4425c2884286a92fde2de6427f_analitic_llm.table
|
||||
4b751a4425c2884286a92fde2de6427f_analitic_ocr.table
|
||||
.vscode/*
|
||||
__pycache__/*
|
||||
|
||||
|
|
@ -0,0 +1,706 @@
|
|||
import fastapi
|
||||
from fastapi import FastAPI, Request
|
||||
from fastapi.responses import HTMLResponse
|
||||
from pydantic import BaseModel
|
||||
import time
|
||||
from fastapi.staticfiles import StaticFiles
|
||||
from fastapi import FastAPI, Query, File, UploadFile
|
||||
#from fastapi.middleware.cors import CORSMiddleware
|
||||
from starlette.middleware.cors import CORSMiddleware
|
||||
import main
|
||||
import os
|
||||
from databases import db
|
||||
import audioread
|
||||
import pandas as pd
|
||||
import statistics
|
||||
pwd = os.getcwd()
|
||||
pathAud="example/audio"
|
||||
pathFact="example/factura"
|
||||
app = FastAPI()
|
||||
#app.mount("/statics", StaticFiles(directory="statics"), name="statics")
|
||||
app.add_middleware(
|
||||
CORSMiddleware,
|
||||
allow_origins=["*"],
|
||||
allow_credentials=True,
|
||||
allow_methods=["*"],
|
||||
allow_headers=["*"],
|
||||
)
|
||||
|
||||
class Response(BaseModel):
|
||||
"""Structure of data to querry of make post from X or article blog
|
||||
"""
|
||||
path: str = Query("", description="Style and sentiments of text")
|
||||
Trusted: str = Query("", description="Style and sentiments of text")
|
||||
model : str = Query("whisper", description="Style and sentiments of text")
|
||||
class Response1(BaseModel):
|
||||
path: str = Query("", description="path file")
|
||||
task_prompt: str = Query("", description="task of model")
|
||||
model: str = Query("", description="model")
|
||||
TrustedOCR: str = Query("", description="truted OCR model")
|
||||
option: str = Query("", description="OCR model option")
|
||||
class Response2(BaseModel):
|
||||
path: str = Query("", description="path file")
|
||||
task_prompt: str = Query("", description="task of model")
|
||||
system: str = Query("", description="prompt system LLM model with ocr and image claude")
|
||||
content: str = Query("%s", description="prompt content LLM model with ocr")
|
||||
max_tokens: int = Query(1024, description="maxtoken LLM OCR model")
|
||||
model: str = Query("Claude-sonnet", description="model")
|
||||
prompt: str = Query("", description="prompt in claude with image")
|
||||
TrustedLLmjson: str = Query("", description="truted OCR model")
|
||||
|
||||
class Response3(BaseModel):
|
||||
"""Structure of data to querry of make post from X or article blog
|
||||
"""
|
||||
path: str = Query("", description="Style and sentiments of text")
|
||||
Trusted: str = Query("", description="Style and sentiments of text")
|
||||
mode : str = Query("whisper", description="Style and sentiments of text")
|
||||
|
||||
|
||||
@app.get("/addTrusted")
|
||||
@app.post("/addTrusted")
|
||||
def addTrusted(response:Response3):
|
||||
path=response.path
|
||||
Trusted=response.Trusted
|
||||
mode=response.mode
|
||||
file_stats = os.stat(path)
|
||||
size=file_stats.st_size / (1024 * 1024)
|
||||
if mode=="voice":
|
||||
with audioread.audio_open(path) as f:
|
||||
duration = f.duration
|
||||
else:
|
||||
duration = 0
|
||||
if db(db.trusted.path == path and db.trusted.mode == mode).count()==0:
|
||||
db.trusted.insert(path=path,trusted=Trusted,mode=mode,size=size,duration =duration )
|
||||
db.commit()
|
||||
return "Add %s in mode %s"%(path,mode)
|
||||
else:
|
||||
db(db.trusted.path == path and db.trusted.mode == mode).update(trusted=Trusted,size=size,duration =duration )
|
||||
db.commit()
|
||||
return "Update %s in mode %s"%(path,mode)
|
||||
|
||||
def list2tablehtml(listdata,model):
|
||||
html="""<h2>Table of {0}</h2>
|
||||
<table style="width:100%">
|
||||
<tr>
|
||||
<th>path</th>
|
||||
<th>time</th>
|
||||
<th>similarity</th>
|
||||
<th>similaritypartial</th>
|
||||
</tr>""".format(model)
|
||||
|
||||
for i in listdata:
|
||||
html=html+""" <tr>
|
||||
<td>%s</td>
|
||||
<td>%s</td>
|
||||
<td>%s</td>
|
||||
<td>%s</td>
|
||||
</tr>
|
||||
"""%(i["path"],i["time"],i["similarity"],i["similaritypartial"])
|
||||
html=html+"""</table>
|
||||
"""
|
||||
return html
|
||||
|
||||
|
||||
def tableVoice(model):
|
||||
rows = db(db.analitic_voice.model==model).select()
|
||||
rows_list = rows.as_list()
|
||||
data=pd.DataFrame(rows_list)
|
||||
durationL=list()
|
||||
for i in rows_list:
|
||||
durationL.append(db(db.trusted.path == i["path"] ).select().last().duration)
|
||||
duration=statistics.mean(durationL)
|
||||
time=pd.pivot_table(data,values=['time','similarity', 'similaritypartial'],index="model")['time'].values[0]
|
||||
similarity=pd.pivot_table(data,values=['time','similarity', 'similaritypartial'],index="model")['similarity'].values[0]
|
||||
similaritypartial=pd.pivot_table(data,values=['time','similarity', 'similaritypartial'],index="model")['similaritypartial'].values[0]
|
||||
efectivetime=time/duration
|
||||
card="""<div class="flip-card">
|
||||
<div class="flip-card-inner">
|
||||
<div class="flip-card-front">
|
||||
<p style="width:300px;height:300px;">{0} </p>
|
||||
</div>
|
||||
<div class="flip-card-back">
|
||||
<h1>time of process (sg)</h1>
|
||||
<p>{1}</p>
|
||||
<h1>similarity</h1>
|
||||
<p>{2}</p>
|
||||
<h1>similaritypartial</h1>
|
||||
<p>{3}</p>
|
||||
<h1>time of audio(sg)</h1>
|
||||
<p>{4}</p>
|
||||
<h1>time in process</h1>
|
||||
<p>{5}</p>
|
||||
</div>
|
||||
</div>
|
||||
</div>""".format(model,time,similarity,similaritypartial,duration,efectivetime)
|
||||
return {"duration":duration,"time":time,"similarity":similarity,"similaritypartial":similaritypartial,"card":card,"data":list2tablehtml(rows_list,model)}
|
||||
|
||||
|
||||
@app.get("/getmetricsvoice")
|
||||
def getMetricsVoice():
|
||||
pass
|
||||
models=list()
|
||||
for row in db().select(db.analitic_voice.model, distinct=True):
|
||||
models.append(row.model)
|
||||
cards=""
|
||||
dataAll=""
|
||||
for model in models:
|
||||
|
||||
Sal=tableVoice(model)
|
||||
cards=cards+Sal["card"]
|
||||
dataAll=dataAll+Sal["data"]
|
||||
|
||||
|
||||
htmlhead="""<!DOCTYPE html>
|
||||
<html lang="es">
|
||||
<head>
|
||||
<meta charset="UTF-8">
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
||||
<title>Evaluacion de modelos voice2txt</title>
|
||||
<style>
|
||||
. container{
|
||||
|
||||
display:flex;
|
||||
|
||||
}
|
||||
/* The flip card container - set the width and height to whatever you want. We have added the border property to demonstrate that the flip itself goes out of the box on hover (remove perspective if you don't want the 3D effect */
|
||||
.flip-card {
|
||||
background-color: transparent;
|
||||
width: 500px;
|
||||
height: 500px;
|
||||
border: 1px solid #f1f1f1;
|
||||
perspective: 1000px; /* Remove this if you don't want the 3D effect */
|
||||
}
|
||||
|
||||
/* This container is needed to position the front and back side */
|
||||
.flip-card-inner {
|
||||
position: relative;
|
||||
width: 100%;
|
||||
height: 100%;
|
||||
text-align: center;
|
||||
transition: transform 0.8s;
|
||||
transform-style: preserve-3d;
|
||||
}
|
||||
|
||||
/* Do an horizontal flip when you move the mouse over the flip box container */
|
||||
.flip-card:hover .flip-card-inner {
|
||||
transform: rotateY(180deg);
|
||||
}
|
||||
|
||||
/* Position the front and back side */
|
||||
.flip-card-front, .flip-card-back {
|
||||
position: absolute;
|
||||
width: 100%;
|
||||
height: 100%;
|
||||
-webkit-backface-visibility: hidden; /* Safari */
|
||||
backface-visibility: hidden;
|
||||
}
|
||||
|
||||
/* Style the front side (fallback if image is missing) */
|
||||
.flip-card-front {
|
||||
background-color: #bbb;
|
||||
color: black;
|
||||
}
|
||||
|
||||
/* Style the back side */
|
||||
.flip-card-back {
|
||||
background-color: dodgerblue;
|
||||
color: white;
|
||||
transform: rotateY(180deg);
|
||||
}
|
||||
</style>
|
||||
</head>"""
|
||||
|
||||
htmlbody="""<body>
|
||||
<h1>Estadisticas modelos de voice</h1>
|
||||
<div class=”container”>
|
||||
{0}
|
||||
</div>
|
||||
{1}
|
||||
</body>
|
||||
</html>
|
||||
""".format(cards,dataAll)
|
||||
html=htmlhead+htmlbody
|
||||
return HTMLResponse(content=html, status_code=200)
|
||||
|
||||
|
||||
@app.get("/EvalVoice")
|
||||
@app.post("/EvalVoice")
|
||||
def EvalVoice(response:Response):
|
||||
path=response.path
|
||||
Trusted=response.Trusted
|
||||
model=response.model
|
||||
if Trusted=="":
|
||||
row=db(db.trusted.path == path and db.trusted.mode == "voice").select().first()
|
||||
try:
|
||||
Trusted=row.trusted
|
||||
except:
|
||||
pass
|
||||
|
||||
if model=="whisper":
|
||||
Sal=main.EvalWhisper(path,Trusted)
|
||||
else:
|
||||
Sal=main.EvalVosk(path,Trusted)
|
||||
if db(db.analitic_voice.path == Sal["path"] and db.analitic_voice.model == Sal["model"]).count()==0:
|
||||
db.analitic_voice.insert(**Sal)
|
||||
db.commit()
|
||||
else:
|
||||
db(db.analitic_voice.path == Sal["path"] and db.analitic_voice.model == Sal["model"]).update(similarity= Sal["similarity"],similaritypartial= Sal["similaritypartial"])
|
||||
db.commit()
|
||||
return Sal
|
||||
|
||||
@app.get("/EvalFact")
|
||||
@app.post("/EvalFact")
|
||||
def EvalFact(response:Response1):
|
||||
path=response.path
|
||||
task_prompt=response.task_prompt
|
||||
option=response.model
|
||||
TrustedOCR=response.TrustedOCR
|
||||
Trusted=TrustedOCR
|
||||
if task_prompt=="":
|
||||
if Trusted=="":
|
||||
row=db(db.trusted.path == path and db.trusted.mode == "OCR").select().first()
|
||||
try:
|
||||
Trusted=row.trusted
|
||||
except:
|
||||
pass
|
||||
Sal=main.EvalFacturas(path,task_prompt,TrustedOCR,option)
|
||||
Sal["path"]=path
|
||||
if db(db.analitic_ocr.path == Sal["path"] and db.analitic_ocr.model == Sal["model"]).count()==0:
|
||||
db.analitic_ocr.insert(**Sal)
|
||||
db.commit()
|
||||
else:
|
||||
db(db.analitic_ocr.path == Sal["path"] and db.analitic_ocr.model == Sal["model"]).update(similarity= Sal["similarity"],similaritypartial= Sal["similaritypartial"],jsonok=Sal["jsonok"])
|
||||
db.commit()
|
||||
|
||||
|
||||
return Sal
|
||||
|
||||
@app.get("/EvalLLMFact")
|
||||
@app.post("/EvalLLMFact")
|
||||
def EvalLLMFact(response:Response2):
|
||||
path=response.path
|
||||
task_prompt=response.task_prompt
|
||||
system=response.system
|
||||
content=response.content
|
||||
max_tokens=response.max_tokens
|
||||
model=response.model
|
||||
prompt=response.prompt
|
||||
TrustedLLmjson=response.TrustedLLmjson
|
||||
|
||||
Sal=main.EvalllmFacturas(path,task_prompt,system,content,max_tokens,model,prompt,TrustedLLmjson)
|
||||
return Sal
|
||||
|
||||
@app.get("/evalvoicehtml")
|
||||
def EvalVoicehtml():
|
||||
dir_list = os.listdir(pathAud)
|
||||
Sal=""
|
||||
t=1
|
||||
for i in dir_list:
|
||||
|
||||
temp="""<option value="%s">Opción %s, %s</option>
|
||||
"""%(str(pwd+"/"+pathAud+"/"+i),str(t),str(i))
|
||||
Sal=Sal+temp
|
||||
t=t+1
|
||||
|
||||
|
||||
html="""<!DOCTYPE html>
|
||||
<html lang="es">
|
||||
<head>
|
||||
<meta charset="UTF-8">
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
||||
<title>Evaluacion de modelos voice2txt</title>
|
||||
<style>
|
||||
body {
|
||||
font-family: Arial, sans-serif;
|
||||
margin: 20px;
|
||||
}
|
||||
input, button {
|
||||
margin: 10px 0;
|
||||
padding: 5px;
|
||||
}
|
||||
#respuesta {
|
||||
margin-top: 20px;
|
||||
padding: 10px;
|
||||
border: 1px solid #ccc;
|
||||
background-color: #f9f9f9;
|
||||
}
|
||||
</style>
|
||||
</head>
|
||||
<body>
|
||||
<h1>Petición POST a API</h1>
|
||||
|
||||
<select id="texto1">
|
||||
%s
|
||||
</select>
|
||||
|
||||
<br>
|
||||
<input type="text" id="texto2" placeholder="Trusted">
|
||||
<br>
|
||||
<select id="texto3">
|
||||
<option value="whisper">whisper</option>
|
||||
<option value="vosk">vosk</option>
|
||||
</select>
|
||||
<br>
|
||||
<button onclick="enviarPeticion()">Enviar petición</button>
|
||||
<div id="respuesta"></div>
|
||||
|
||||
<script>
|
||||
function enviarPeticion() {
|
||||
const texto1 = document.getElementById('texto1').value;
|
||||
const texto2 = document.getElementById('texto2').value;
|
||||
const texto3 = document.getElementById('texto3').value;
|
||||
const datos = {
|
||||
path: texto1,
|
||||
Trusted: texto2,
|
||||
model: texto3
|
||||
};
|
||||
|
||||
fetch('/EvalVoice', {
|
||||
method: 'POST',
|
||||
headers: {
|
||||
'Content-Type': 'application/json'
|
||||
},
|
||||
body: JSON.stringify(datos)
|
||||
})
|
||||
.then(response => response.json())
|
||||
.then(data => {
|
||||
document.getElementById('respuesta').innerHTML = JSON.stringify(data, null, 2);
|
||||
})
|
||||
.catch(error => {
|
||||
document.getElementById('respuesta').innerHTML = 'Error: ' + error;
|
||||
});
|
||||
}
|
||||
</script>
|
||||
</body>
|
||||
</html>
|
||||
"""%(Sal)
|
||||
return HTMLResponse(content=html, status_code=200)
|
||||
|
||||
|
||||
@app.get("/evalocrfactura")
|
||||
def EvalOCRFactura():
|
||||
dir_list = os.listdir(pathFact)
|
||||
Sal=""
|
||||
t=1
|
||||
for i in dir_list:
|
||||
temp="""<option value="%s">Opción %s, %s</option>
|
||||
"""%(str(pwd+"/"+pathFact+"/"+i),str(t),str(i))
|
||||
Sal=Sal+temp
|
||||
t=t+1
|
||||
html="""<!DOCTYPE html>
|
||||
<html lang="es">
|
||||
<head>
|
||||
<meta charset="UTF-8">
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
||||
<title>Evaluacion de modelos OCR</title>
|
||||
<style>
|
||||
body {
|
||||
font-family: Arial, sans-serif;
|
||||
margin: 20px;
|
||||
}
|
||||
input, button {
|
||||
margin: 10px 0;
|
||||
padding: 5px;
|
||||
}
|
||||
#respuesta {
|
||||
margin-top: 20px;
|
||||
padding: 10px;
|
||||
border: 1px solid #ccc;
|
||||
background-color: #f9f9f9;
|
||||
}
|
||||
</style>
|
||||
</head>
|
||||
<body>
|
||||
<h1>Petición POST a API</h1>
|
||||
<select id="texto1">
|
||||
%s
|
||||
</select>
|
||||
<br>
|
||||
|
||||
<select id="texto2">
|
||||
<option value="More Detailed Caption">More Detailed Caption</option>
|
||||
<option value="OCR">OCR</option>
|
||||
<option value="parsed">parsed</option>
|
||||
<option value="scan">scan</option>
|
||||
</select>
|
||||
<br>
|
||||
<input type="text" id="texto3" placeholder="TrustedOCR">
|
||||
<br>
|
||||
<input type="text" id="texto4" placeholder="option">
|
||||
<br>
|
||||
<button onclick="enviarPeticion()">Enviar petición</button>
|
||||
<div id="respuesta"></div>
|
||||
|
||||
<script>
|
||||
function enviarPeticion() {
|
||||
const texto1 = document.getElementById('texto1').value;
|
||||
const texto2 = document.getElementById('texto2').value;
|
||||
const texto3 = document.getElementById('texto3').value;
|
||||
const texto4 = document.getElementById('texto4').value;
|
||||
const datos = {
|
||||
path: texto1,
|
||||
task_prompt: texto2,
|
||||
TrustedOCR: texto3,
|
||||
option: texto4
|
||||
};
|
||||
|
||||
fetch('/EvalFact', {
|
||||
method: 'POST',
|
||||
headers: {
|
||||
'Content-Type': 'application/json'
|
||||
},
|
||||
body: JSON.stringify(datos)
|
||||
})
|
||||
.then(response => response.json())
|
||||
.then(data => {
|
||||
document.getElementById('respuesta').innerHTML = JSON.stringify(data, null, 2);
|
||||
})
|
||||
.catch(error => {
|
||||
document.getElementById('respuesta').innerHTML = 'Error: ' + error;
|
||||
});
|
||||
}
|
||||
</script>
|
||||
</body>
|
||||
</html>
|
||||
"""%(Sal)
|
||||
return HTMLResponse(content=html, status_code=200)
|
||||
|
||||
def list2tablehtmlOCR(listdata,model):
|
||||
html="""<h2>Table of {0}</h2>
|
||||
<table style="width:100%">
|
||||
<tr>
|
||||
<th>path</th>
|
||||
<th>time</th>
|
||||
<th>similarity</th>
|
||||
<th>similaritypartial</th>
|
||||
</tr>""".format(model)
|
||||
|
||||
for i in listdata:
|
||||
html=html+""" <tr>
|
||||
<td>%s</td>
|
||||
<td>%s</td>
|
||||
<td>%s</td>
|
||||
<td>%s</td>
|
||||
</tr>
|
||||
"""%(i["path"],i["time"],i["similarity"],i["similaritypartial"])
|
||||
html=html+"""</table>
|
||||
"""
|
||||
return html
|
||||
|
||||
|
||||
def tableOCR(model):
|
||||
rows = db(db.analitic_ocr.model==model).select()
|
||||
rows_list = rows.as_list()
|
||||
data=pd.DataFrame(rows_list)
|
||||
time=pd.pivot_table(data,values=['time','similarity', 'similaritypartial'],index="model")['time'].values[0]
|
||||
similarity=pd.pivot_table(data,values=['time','similarity', 'similaritypartial'],index="model")['similarity'].values[0]
|
||||
similaritypartial=pd.pivot_table(data,values=['time','similarity', 'similaritypartial'],index="model")['similaritypartial'].values[0]
|
||||
card="""<div class="flip-card">
|
||||
<div class="flip-card-inner">
|
||||
<div class="flip-card-front">
|
||||
<p style="width:300px;height:300px;">{0} </p>
|
||||
</div>
|
||||
<div class="flip-card-back">
|
||||
<h1>time of process (sg)</h1>
|
||||
<p>{1}</p>
|
||||
<h1>similarity</h1>
|
||||
<p>{2}</p>
|
||||
<h1>similaritypartial</h1>
|
||||
<p>{3}</p>
|
||||
</div>
|
||||
</div>
|
||||
</div>""".format(model,time,similarity,similaritypartial)
|
||||
return {"time":time,"similarity":similarity,"similaritypartial":similaritypartial,"card":card,"data":list2tablehtmlOCR(rows_list,model)}
|
||||
|
||||
|
||||
|
||||
@app.get("/getmetricsocr")
|
||||
def getMetricsOCR():
|
||||
models=list()
|
||||
for row in db().select(db.analitic_ocr.model, distinct=True):
|
||||
models.append(row.model)
|
||||
cards=""
|
||||
dataAll=""
|
||||
for model in models:
|
||||
Sal=tableOCR(model)
|
||||
cards=cards+Sal["card"]
|
||||
dataAll=dataAll+Sal["data"]
|
||||
htmlhead="""<!DOCTYPE html>
|
||||
<html lang="es">
|
||||
<head>
|
||||
<meta charset="UTF-8">
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
||||
<title>Evaluacion de modelos voice2txt</title>
|
||||
<style>
|
||||
. container{
|
||||
|
||||
display:flex;
|
||||
|
||||
}
|
||||
/* The flip card container - set the width and height to whatever you want. We have added the border property to demonstrate that the flip itself goes out of the box on hover (remove perspective if you don't want the 3D effect */
|
||||
.flip-card {
|
||||
background-color: transparent;
|
||||
width: 500px;
|
||||
height: 500px;
|
||||
border: 1px solid #f1f1f1;
|
||||
perspective: 1000px; /* Remove this if you don't want the 3D effect */
|
||||
}
|
||||
|
||||
/* This container is needed to position the front and back side */
|
||||
.flip-card-inner {
|
||||
position: relative;
|
||||
width: 100%;
|
||||
height: 100%;
|
||||
text-align: center;
|
||||
transition: transform 0.8s;
|
||||
transform-style: preserve-3d;
|
||||
}
|
||||
|
||||
/* Do an horizontal flip when you move the mouse over the flip box container */
|
||||
.flip-card:hover .flip-card-inner {
|
||||
transform: rotateY(180deg);
|
||||
}
|
||||
|
||||
/* Position the front and back side */
|
||||
.flip-card-front, .flip-card-back {
|
||||
position: absolute;
|
||||
width: 100%;
|
||||
height: 100%;
|
||||
-webkit-backface-visibility: hidden; /* Safari */
|
||||
backface-visibility: hidden;
|
||||
}
|
||||
|
||||
/* Style the front side (fallback if image is missing) */
|
||||
.flip-card-front {
|
||||
background-color: #bbb;
|
||||
color: black;
|
||||
}
|
||||
|
||||
/* Style the back side */
|
||||
.flip-card-back {
|
||||
background-color: dodgerblue;
|
||||
color: white;
|
||||
transform: rotateY(180deg);
|
||||
}
|
||||
</style>
|
||||
</head>"""
|
||||
|
||||
htmlbody="""<body>
|
||||
<h1>Estadisticas modelos de OCR</h1>
|
||||
<div class=”container”>
|
||||
{0}
|
||||
</div>
|
||||
{1}
|
||||
</body>
|
||||
</html>
|
||||
""".format(cards,dataAll)
|
||||
html=htmlhead+htmlbody
|
||||
return HTMLResponse(content=html, status_code=200)
|
||||
|
||||
|
||||
|
||||
@app.get("/evalllmfacturas")
|
||||
def EvalllmFacturas():
|
||||
dir_list = os.listdir(pathFact)
|
||||
Sal=""
|
||||
t=1
|
||||
for i in dir_list:
|
||||
temp="""<option value="%s">Opción %s, %s</option>
|
||||
"""%(str(pwd+"/"+pathFact+"/"+i),str(t),str(i))
|
||||
Sal=Sal+temp
|
||||
t=t+1
|
||||
html="""<!DOCTYPE html>
|
||||
<html lang="es">
|
||||
<head>
|
||||
<meta charset="UTF-8">
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
||||
<title>Evaluacion modelos LLM</title>
|
||||
<style>
|
||||
body {
|
||||
font-family: Arial, sans-serif;
|
||||
margin: 20px;
|
||||
}
|
||||
input, button {
|
||||
margin: 10px 0;
|
||||
padding: 5px;
|
||||
}
|
||||
#respuesta {
|
||||
margin-top: 20px;
|
||||
padding: 10px;
|
||||
border: 1px solid #ccc;
|
||||
background-color: #f9f9f9;
|
||||
}
|
||||
</style>
|
||||
</head>
|
||||
<body>
|
||||
<h1>Petición POST a API</h1>
|
||||
<select id="texto1">
|
||||
%s
|
||||
</select>
|
||||
<br>
|
||||
|
||||
<select id="texto2">
|
||||
<option value="">N.A.</option>
|
||||
<option value="More Detailed Caption">More Detailed Caption</option>
|
||||
<option value="OCR">OCR</option>
|
||||
<option value="parsed">parsed</option>
|
||||
<option value="scan">scan</option>
|
||||
</select>
|
||||
<br>
|
||||
<input type="text" id="texto3" placeholder="system" value="Eres un chatbot amable">
|
||||
<br>
|
||||
<input type="text" id="texto4" placeholder="content" value="%s">
|
||||
<br>
|
||||
<input type="number" id="texto5" placeholder="max_tokens" value=1024>
|
||||
<br>
|
||||
<input type="text" id="texto6" placeholder="model" value="Claude-sonnet">
|
||||
<br>
|
||||
<input type="text" id="texto7" placeholder="prompt" value="Analiza la factura">
|
||||
<br>
|
||||
<input type="text" id="texto8" placeholder="TrustedLLmjson" value="{'A':''}">
|
||||
<br>
|
||||
<button onclick="enviarPeticion()">Enviar petición</button>
|
||||
<div id="respuesta"></div>
|
||||
|
||||
<script>
|
||||
function enviarPeticion() {
|
||||
const texto1 = document.getElementById('texto1').value;
|
||||
const texto2 = document.getElementById('texto2').value;
|
||||
const texto3 = document.getElementById('texto3').value;
|
||||
const texto4 = document.getElementById('texto4').value;
|
||||
const texto5 = document.getElementById('texto5').value;
|
||||
const texto6 = document.getElementById('texto6').value;
|
||||
const texto7 = document.getElementById('texto7').value;
|
||||
const texto8 = document.getElementById('texto8').value;
|
||||
|
||||
const datos = {
|
||||
path: texto1,
|
||||
task_prompt: texto2,
|
||||
system: texto3,
|
||||
content:texto4,
|
||||
max_tokens:texto5,
|
||||
model:texto6,
|
||||
prompt:texto7,
|
||||
TrustedLLmjson:texto8,
|
||||
};
|
||||
|
||||
fetch('/EvalLLMFact', {
|
||||
method: 'POST',
|
||||
headers: {
|
||||
'Content-Type': 'application/json'
|
||||
},
|
||||
body: JSON.stringify(datos)
|
||||
})
|
||||
.then(response => response.json())
|
||||
.then(data => {
|
||||
document.getElementById('respuesta').innerHTML = JSON.stringify(data, null, 2);
|
||||
})
|
||||
.catch(error => {
|
||||
document.getElementById('respuesta').innerHTML = 'Error: ' + error;
|
||||
});
|
||||
}
|
||||
</script>
|
||||
</body>
|
||||
</html>
|
||||
"""%(Sal,"%s")
|
||||
return HTMLResponse(content=html, status_code=200)
|
||||
|
|
@ -0,0 +1,43 @@
|
|||
from pydal import DAL, Field
|
||||
db = DAL("sqlite://databases/storage.db")
|
||||
db.define_table(
|
||||
"trusted",
|
||||
Field("path"),
|
||||
Field("mode"),
|
||||
Field("trusted"),
|
||||
Field("duration",type="double"),
|
||||
Field("size",type="double")
|
||||
)
|
||||
db.define_table(
|
||||
"analitic_voice",
|
||||
Field("content"),
|
||||
Field("trusted"),
|
||||
Field("model"),
|
||||
Field("time", type="double"),
|
||||
Field("path"),
|
||||
Field("similarity", type="double"),
|
||||
Field("similaritypartial", type="double")
|
||||
)
|
||||
|
||||
db.define_table(
|
||||
"analitic_ocr",
|
||||
Field("content"),
|
||||
Field("trusted"),
|
||||
Field("model"),
|
||||
Field("time", type="double"),
|
||||
Field("path"),
|
||||
Field("similarity", type="double"),
|
||||
Field("similaritypartial", type="double"),
|
||||
Field("jsonok" ,type="integer")
|
||||
)
|
||||
|
||||
db.define_table(
|
||||
"analitic_llm",
|
||||
Field("content"),
|
||||
Field("trusted"),
|
||||
Field("model"),
|
||||
Field("time", type="double"),
|
||||
Field("path"),
|
||||
Field("similarity", type="double"),
|
||||
Field("similaritypartial", type="double")
|
||||
)
|
||||
|
|
@ -0,0 +1,157 @@
|
|||
import requests
|
||||
import evaluate
|
||||
import deepdiff
|
||||
import json
|
||||
from fuzzywuzzy import fuzz
|
||||
from deepdiff import DeepDiff
|
||||
from deepdiff import Delta
|
||||
import databases
|
||||
#print(evaluate.list_evaluation_modules())
|
||||
urlAud="http://127.0.0.1:7870/"
|
||||
urlText="http://127.0.0.1:7869"
|
||||
password="1223Aer*"
|
||||
def EvalVoice2Text(endpoint,datajson,Trusted):
|
||||
"""Evaluate Voice 2 text
|
||||
"""
|
||||
apiUrl=urlAud+endpoint
|
||||
response = requests.get(apiUrl, json=datajson)
|
||||
print(datajson)
|
||||
A=json.loads(response.content)
|
||||
print(A)
|
||||
time=A['time']
|
||||
|
||||
similarity=fuzz.ratio( Trusted.strip().lower(),A['message'].strip().lower())
|
||||
similarityPartial=fuzz.partial_ratio( Trusted.strip().lower(),A['message'].strip().lower())
|
||||
path=datajson["local"]
|
||||
model=datajson["model"]
|
||||
message=A['message']
|
||||
return {"content":message,
|
||||
"trusted":Trusted,
|
||||
"model":model,
|
||||
"time":time,
|
||||
"similarity":similarity,
|
||||
"similaritypartial":similarityPartial,
|
||||
"path":path
|
||||
}
|
||||
|
||||
|
||||
def EvalWhisper(path,Trusted=""):
|
||||
endpoint="/voice2txt"
|
||||
datajson={"url":"","password":password ,"model":"whisper","local":path}
|
||||
return EvalVoice2Text(endpoint,datajson,Trusted)
|
||||
|
||||
|
||||
# EvalWhisper(path="example/AwACAgEAAxkBAAIBw2YX8o2vGGCNtZCXk7mY1Bm5w__lAAJmBAACxe7ARI1fUWAGcz_RNAQ.ogg",
|
||||
# Trusted="Hoy compre dos medicinas Tereleji en Cruz Verde por un monto de 494 mil 400 pesos colombianos.",
|
||||
# endpoint="/voice2txt")
|
||||
|
||||
def EvalVosk(path,Trusted=""):
|
||||
endpoint="/voice2txtlocal"
|
||||
datajson={"url":"","password":password ,"model":"models/vosk-model-small-es-0.42","local":path}
|
||||
return EvalVoice2Text(endpoint,datajson,Trusted)
|
||||
|
||||
|
||||
|
||||
# EvalVosk(path="example/AwACAgEAAxkBAAIBw2YX8o2vGGCNtZCXk7mY1Bm5w__lAAJmBAACxe7ARI1fUWAGcz_RNAQ.ogg",
|
||||
# Trusted="Hoy compre dos medicinas Tereleji en Cruz Verde por un monto de 494 mil 400 pesos colombianos.",
|
||||
# endpoint="/voice2txtlocal")
|
||||
|
||||
|
||||
def ocrfacturas(path,task_prompt):
|
||||
apiUrl=urlText+'/parsedimage3'
|
||||
datajson={"path":path,"task_prompt":task_prompt,"password":password}
|
||||
response = requests.get(apiUrl, json=datajson)
|
||||
return response.content
|
||||
|
||||
def llmFacturas(path,task_prompt,system,content,max_tokens,model):
|
||||
apiUrl=urlText+'/parsedimage4'
|
||||
datajson={"path":path,"task_prompt":task_prompt,"system":system,"content":content,"max_tokens":max_tokens,"model":model,"password":password}
|
||||
response = requests.get(apiUrl, json=datajson)
|
||||
return response.content
|
||||
|
||||
def llmFacturas2(path,prompt,system,model):
|
||||
apiUrl=urlText+'/parsedimage2'
|
||||
datajson={"path":path,"prompt":prompt,"system":system,"model":model,"password":password}
|
||||
response = requests.get(apiUrl, json=datajson)
|
||||
return response.content
|
||||
|
||||
def EvalParsedImage(path="/home/mario/Repositorios/EvalDataSetHugging/example/Gmail/20240530_112812.jpg"):
|
||||
endpoint="/parsedimage"
|
||||
jsonT={"path":path,"password":password}
|
||||
response=requests.get(urlText+endpoint,json=jsonT)
|
||||
return response.content
|
||||
|
||||
def EvalParsedImage5(path="/home/mario/Repositorios/EvalDataSetHugging/example/Gmail/20240530_112812.jpg",option="teserac"):
|
||||
endpoint="/parsedimage5"
|
||||
jsonT={"path":path,"password":password,"option":option}
|
||||
response=requests.get(urlText+endpoint,json=jsonT)
|
||||
return response.content
|
||||
|
||||
def EvalFacturas(path,task_prompt,TrustedOCR,option=""):
|
||||
if task_prompt=="parsed":
|
||||
OCR=EvalParsedImage(path)
|
||||
if task_prompt=="More Detailed Caption" or task_prompt=='OCR':
|
||||
OCR=ocrfacturas(path,task_prompt)
|
||||
if task_prompt=="scan":
|
||||
OCR=EvalParsedImage5(path,option)
|
||||
model=json.loads(OCR)["model"]
|
||||
content=json.loads(OCR)["content"]
|
||||
time=json.loads(OCR)["time"]
|
||||
try:
|
||||
TrustedOCR=json.loads(TrustedOCR)
|
||||
jsonok=1
|
||||
except:
|
||||
jsonok=0
|
||||
pass
|
||||
similarity=fuzz.ratio( str(TrustedOCR).strip().lower(),str(content).strip().lower())
|
||||
similarityPartial=fuzz.partial_ratio( str(TrustedOCR).strip().lower(),str(content).strip().lower())
|
||||
return {"content":content,
|
||||
"trusted":TrustedOCR,
|
||||
"similarity":similarity,
|
||||
"similaritypartial":similarityPartial,
|
||||
"model":model,
|
||||
"time":time,
|
||||
"jsonok":jsonok
|
||||
}
|
||||
def changemodel(model):
|
||||
if model=="Claude-sonnet":
|
||||
model="claude-3-5-sonnet-20240620"
|
||||
elif model=="Claude-opus":
|
||||
model="claude-3-opus-20240229"
|
||||
elif model=="Claude-haiku":
|
||||
model="claude-3-haiku-20240307"
|
||||
return model
|
||||
|
||||
def EvalllmFacturas(path,task_prompt,system,content,max_tokens,model,prompt,TrustedLLmjson):
|
||||
model=changemodel(model)
|
||||
if model.count("claude")>0 and task_prompt=="":
|
||||
LLmjson=llmFacturas2(path=path,prompt=prompt,system=system,model=model)
|
||||
else:
|
||||
LLmjson=llmFacturas(path=path,task_prompt=task_prompt,system=system,content=content,max_tokens=max_tokens,model=model)
|
||||
TrustedLLmjson=json.loads(TrustedLLmjson)
|
||||
return {"content":LLmjson,"trusted":TrustedLLmjson}
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
#EvalFacturas(path="example/Factura2.jpg",task_prompt="OCR",system="",content="Analiza el siguiente texto: %s",max_tokens=200,model="claude-sonnet")
|
||||
|
||||
def EvalClassImage(path="/home/mario/Repositorios/EvalDataSetHugging/example/Gmail/20240530_112812.jpg",):
|
||||
endpoint="classificateimage"
|
||||
jsonT={"path":path,"password":password}
|
||||
response=requests.get(urlText+endpoint,json=jsonT)
|
||||
print(response.content)
|
||||
|
||||
#To Do
|
||||
def EvalGeneratedText(prompt="",model="",):
|
||||
pass
|
||||
|
||||
def EvalGenerateVoice():
|
||||
def GenerateVoice():
|
||||
pass
|
||||
def Voice2txt():
|
||||
pass
|
||||
|
||||
|
||||
|
||||
Loading…
Reference in New Issue