diff --git a/Cargo.toml b/Cargo.toml
index b324cf36246dea409ec30ad87557f18a9c87224d..2a08cd07d68d3f8392ccf7e06293602731b31ed8 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -32,4 +32,4 @@ regex = "1.10.2"
name = "doc_gpt"
[workspace]
-members = [".", "migration"]
\ No newline at end of file
+members = [".", "migration"]
diff --git a/migration/src/m20220101_000001_create_table.rs b/migration/src/m20220101_000001_create_table.rs
index 41287e8d11abe9b9173963ba0b9e2fc8e74e8e52..e6892d7e1383510924f36a462c4a7c034332ef34 100644
--- a/migration/src/m20220101_000001_create_table.rs
+++ b/migration/src/m20220101_000001_create_table.rs
@@ -201,7 +201,7 @@ impl MigrationTrait for Migration {
.col(ColumnDef::new(DocInfo::Location).string().not_null())
.col(ColumnDef::new(DocInfo::Size).big_integer().not_null())
.col(ColumnDef::new(DocInfo::Type).string().not_null())
- .col(ColumnDef::new(DocInfo::ThumbnailBase64).string().not_null())
+ .col(ColumnDef::new(DocInfo::ThumbnailBase64).string().default(""))
.comment("doc type|folder")
.col(
ColumnDef::new(DocInfo::CreatedAt)
@@ -274,28 +274,28 @@ impl MigrationTrait for Migration {
.values_panic([
(1).into(),
"Video".into(),
- ".*\\.(mpg|mpeg|avi|rm|rmvb|mov|wmv|asf|dat|asx|wvx|mpe|mpa)".into(),
+ ".*\\.(mpg|mpeg|avi|rm|rmvb|mov|wmv|asf|dat|asx|wvx|mpe|mpa|mp4)".into(),
(1).into(),
(1).into(),
])
.values_panic([
(1).into(),
"Picture".into(),
- ".*\\.(jpg|jpeg|png|tif|gif|pcx|tga|exif|fpx|svg|psd|cdr|pcd|dxf|ufo|eps|ai|raw|WMF|webp|avif|apng)".into(),
+ ".*\\.(jpg|jpeg|png|tif|gif|pcx|tga|exif|fpx|svg|psd|cdr|pcd|dxf|ufo|eps|ai|raw|WMF|webp|avif|apng|icon|ico)".into(),
(2).into(),
(2).into(),
])
.values_panic([
(1).into(),
"Music".into(),
- ".*\\.(WAV|FLAC|APE|ALAC|WavPack|WV|MP3|AAC|Ogg|Vorbis|Opus)".into(),
+ ".*\\.(wav|flac|ape|alac|wavpack|wv|mp3|aac|ogg|vorbis|opus|mp3)".into(),
(3).into(),
(3).into(),
])
.values_panic([
(1).into(),
"Document".into(),
- ".*\\.(pdf|doc|ppt|yml|xml|htm|json|csv|txt|ini|xsl|wps|rtf|hlp)".into(),
+ ".*\\.(pdf|doc|ppt|yml|xml|htm|json|csv|txt|ini|xsl|wps|rtf|hlp|pages|numbers|key)".into(),
(3).into(),
(3).into(),
])
diff --git a/python/svr/add_thumbnail2file.py b/python/svr/add_thumbnail2file.py
new file mode 100644
index 0000000000000000000000000000000000000000..e4558ca9a2f601b4cced9fd473815e6653a817e8
--- /dev/null
+++ b/python/svr/add_thumbnail2file.py
@@ -0,0 +1,118 @@
+import sys, datetime, random, re, cv2
+from os.path import dirname, realpath
+sys.path.append(dirname(realpath(__file__)) + "/../")
+from util.db_conn import Postgres
+from util.minio_conn import HuMinio
+from util import findMaxDt
+import base64
+from io import BytesIO
+import pandas as pd
+from PIL import Image
+import pdfplumber
+
+
+PG = Postgres("infiniflow", "docgpt")
+MINIO = HuMinio("infiniflow")
+def set_thumbnail(did, base64):
+ sql = f"""
+ update doc_info set thumbnail_base64='{base64}'
+ where
+ did={did}
+ """
+ PG.update(sql)
+
+
+def collect(comm, mod, tm):
+ sql = f"""
+ select
+ did, uid, doc_name, location, updated_at
+ from doc_info
+ where
+ updated_at >= '{tm}'
+ and MOD(did, {comm}) = {mod}
+ and is_deleted=false
+ and type <> 'folder'
+ and thumbnail_base64=''
+ order by updated_at asc
+ limit 10
+ """
+ docs = PG.select(sql)
+ if len(docs) == 0:return pd.DataFrame()
+
+ mtm = str(docs["updated_at"].max())[:19]
+ print("TOTAL:", len(docs), "To: ", mtm)
+ return docs
+
+
+def build(row):
+ if not re.search(r"\.(pdf|jpg|jpeg|png|gif|svg|apng|icon|ico|webp|mpg|mpeg|avi|rm|rmvb|mov|wmv|mp4)$",
+ row["doc_name"].lower().strip()):
+ set_thumbnail(row["did"], "_")
+ return
+
+ def thumbnail(img, SIZE=128):
+ w,h = img.size
+ p = SIZE/max(w, h)
+ w, h = int(w*p), int(h*p)
+ img.thumbnail((w, h))
+ buffered = BytesIO()
+ try:
+ img.save(buffered, format="JPEG")
+ except Exception as e:
+ try:
+ img.save(buffered, format="PNG")
+ except Exception as ee:
+ pass
+ return base64.b64encode(buffered.getvalue()).decode("utf-8")
+
+
+ iobytes = BytesIO(MINIO.get("%s-upload"%str(row["uid"]), row["location"]))
+ if re.search(r"\.pdf$", row["doc_name"].lower().strip()):
+ pdf = pdfplumber.open(iobytes)
+ img = pdf.pages[0].to_image().annotated
+ set_thumbnail(row["did"], thumbnail(img))
+
+ if re.search(r"\.(jpg|jpeg|png|gif|svg|apng|webp|icon|ico)$", row["doc_name"].lower().strip()):
+ img = Image.open(iobytes)
+ set_thumbnail(row["did"], thumbnail(img))
+
+ if re.search(r"\.(mpg|mpeg|avi|rm|rmvb|mov|wmv|mp4)$", row["doc_name"].lower().strip()):
+ url = MINIO.get_presigned_url("%s-upload"%str(row["uid"]),
+ row["location"],
+ expires=datetime.timedelta(seconds=60)
+ )
+ cap = cv2.VideoCapture(url)
+ succ = cap.isOpened()
+ i = random.randint(1, 11)
+ while succ:
+ ret, frame = cap.read()
+ if not ret: break
+ if i > 0:
+ i -= 1
+ continue
+ img = Image.fromarray(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))
+ print(img.size)
+ set_thumbnail(row["did"], thumbnail(img))
+ cap.release()
+ cv2.destroyAllWindows()
+
+
+def main(comm, mod):
+ global model
+ tm_fnm = f"res/thumbnail-{comm}-{mod}.tm"
+ tm = findMaxDt(tm_fnm)
+ rows = collect(comm, mod, tm)
+ if len(rows) == 0:return
+
+ tmf = open(tm_fnm, "a+")
+ for _, r in rows.iterrows():
+ build(r)
+ tmf.write(str(r["updated_at"]) + "\n")
+ tmf.close()
+
+
+if __name__ == "__main__":
+ from mpi4py import MPI
+ comm = MPI.COMM_WORLD
+ main(comm.Get_size(), comm.Get_rank())
+
diff --git a/python/util/minio_conn.py b/python/util/minio_conn.py
index a8e20bac7f96a8d234a3e417855bb28445504c15..05875ac515aaf2b9f3c4416b202cc4ebcb2361c5 100644
--- a/python/util/minio_conn.py
+++ b/python/util/minio_conn.py
@@ -54,11 +54,24 @@ class HuMinio(object):
r = self.conn.get_object(bucket, fnm)
return r.read()
except Exception as e:
- logging.error(f"Fail get {bucket}/{fnm}: "+str(e))
+ logging.error(f"fail get {bucket}/{fnm}: "+str(e))
self.__open__()
time.sleep(1)
return
+
+ def get_presigned_url(self, bucket, fnm, expires):
+ for _ in range(10):
+ try:
+ return self.conn.get_presigned_url("GET", bucket, fnm, expires)
+ except Exception as e:
+ logging.error(f"fail get {bucket}/{fnm}: "+str(e))
+ self.__open__()
+ time.sleep(1)
+ return
+
+
+
if __name__ == "__main__":
conn = HuMinio("infiniflow")
fnm = "/opt/home/kevinhu/docgpt/upload/13/11-408.jpg"
diff --git a/src/api/doc_info.rs b/src/api/doc_info.rs
index 9169bf0b84f3667699c68301ee23d2896f8a2e45..af1ee131679fb3b705def8927d75f57699f0f647 100644
--- a/src/api/doc_info.rs
+++ b/src/api/doc_info.rs
@@ -1,6 +1,7 @@
-use std::collections::HashMap;
+use std::collections::{HashMap};
use std::io::BufReader;
use actix_multipart_extract::{ File, Multipart, MultipartForm };
+use actix_web::web::Bytes;
use actix_web::{ HttpResponse, post, web };
use chrono::{ Utc, FixedOffset };
use minio::s3::args::{ BucketExistsArgs, MakeBucketArgs, PutObjectArgs };
@@ -68,7 +69,7 @@ pub struct UploadForm {
fn file_type(filename: &String) -> String {
let fnm = filename.to_lowercase();
if
- let Some(_) = Regex::new(r"\.(mpg|mpeg|avi|rm|rmvb|mov|wmv|asf|dat|asx|wvx|mpe|mpa)$")
+ let Some(_) = Regex::new(r"\.(mpg|mpeg|avi|rm|rmvb|mov|wmv|asf|dat|asx|wvx|mpe|mpa|mp4)$")
.unwrap()
.captures(&fnm)
{
@@ -76,7 +77,7 @@ fn file_type(filename: &String) -> String {
}
if
let Some(_) = Regex::new(
- r"\.(jpg|jpeg|png|tif|gif|pcx|tga|exif|fpx|svg|psd|cdr|pcd|dxf|ufo|eps|ai|raw|WMF|webp|avif|apng)$"
+ r"\.(jpg|jpeg|png|tif|gif|pcx|tga|exif|fpx|svg|psd|cdr|pcd|dxf|ufo|eps|ai|raw|WMF|webp|avif|apng|icon|ico)$"
)
.unwrap()
.captures(&fnm)
@@ -84,14 +85,14 @@ fn file_type(filename: &String) -> String {
return "Picture".to_owned();
}
if
- let Some(_) = Regex::new(r"\.(WAV|FLAC|APE|ALAC|WavPack|WV|MP3|AAC|Ogg|Vorbis|Opus)$")
+ let Some(_) = Regex::new(r"\.(wav|flac|ape|alac|wavpack|wv|mp3|aac|ogg|vorbis|opus|mp3)$")
.unwrap()
.captures(&fnm)
{
return "Music".to_owned();
}
if
- let Some(_) = Regex::new(r"\.(pdf|doc|ppt|yml|xml|htm|json|csv|txt|ini|xsl|wps|rtf|hlp)$")
+ let Some(_) = Regex::new(r"\.(pdf|doc|ppt|yml|xml|htm|json|csv|txt|ini|xsl|wps|rtf|hlp|pages|numbers|key)$")
.unwrap()
.captures(&fnm)
{
@@ -100,6 +101,7 @@ fn file_type(filename: &String) -> String {
"Other".to_owned()
}
+
#[post("/v1.0/upload")]
async fn upload(
payload: Multipart<UploadForm>,