litdb 2.1.2__tar.gz → 2.1.4__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- {litdb-2.1.2 → litdb-2.1.4}/PKG-INFO +6 -1
- {litdb-2.1.2 → litdb-2.1.4}/README.org +2 -0
- {litdb-2.1.2 → litdb-2.1.4}/pyproject.toml +7 -2
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/chat.py +9 -6
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/cli.py +20 -2
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/db.py +20 -1
- litdb-2.1.4/src/litdb/images.py +73 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb.egg-info/PKG-INFO +6 -1
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb.egg-info/SOURCES.txt +1 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb.egg-info/requires.txt +3 -0
- {litdb-2.1.2 → litdb-2.1.4}/setup.cfg +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/__init__.py +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/app.py +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/audio.py +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/bibtex.py +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/coa.py +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/gpt.py +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/lab.py +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/openalex.py +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/pdf.py +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/utils.py +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb/youtube.py +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb.egg-info/dependency_links.txt +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb.egg-info/entry_points.txt +0 -0
- {litdb-2.1.2 → litdb-2.1.4}/src/litdb.egg-info/top_level.txt +0 -0
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
Metadata-Version: 2.2
|
|
2
2
|
Name: litdb
|
|
3
|
-
Version: 2.1.
|
|
3
|
+
Version: 2.1.4
|
|
4
4
|
Summary: A literature database tool with GPT integration.
|
|
5
5
|
Author-email: John Kitchin <jkitchin@andrew.cmu.edu>
|
|
6
6
|
Maintainer-email: John Kitchin <jkitchin@andrew.cmu.edu>
|
|
@@ -40,6 +40,9 @@ Requires-Dist: IPython
|
|
|
40
40
|
Requires-Dist: yt_dlp
|
|
41
41
|
Requires-Dist: webvtt-py
|
|
42
42
|
Requires-Dist: streamlit
|
|
43
|
+
Requires-Dist: pyperclip
|
|
44
|
+
Requires-Dist: PIL
|
|
45
|
+
Requires-Dist: pillow_heif
|
|
43
46
|
|
|
44
47
|
#+title: litdb - a literature and document database
|
|
45
48
|
|
|
@@ -61,6 +64,8 @@ The idea is you add papers to your database, and then you can search it with nat
|
|
|
61
64
|
5. https://youtube.com/live/j7rItPwWDaY litdb and Jupyter Lab
|
|
62
65
|
6. https://youtube.com/live/SUtvtc7l6y0 litdb + GPT enhancements
|
|
63
66
|
7. https://youtube.com/live/3FZ1ROnCC6Y litdb + LiteLLM and streamlit
|
|
67
|
+
8. https://www.youtube.com/live/IKKTQSTXQmc litdb + Youtube and audio
|
|
68
|
+
9. https://youtube.com/live/MEf9rPI0Z1M litdb + Image search with text and image queries using CLIP
|
|
64
69
|
|
|
65
70
|
** installation
|
|
66
71
|
|
|
@@ -18,6 +18,8 @@ The idea is you add papers to your database, and then you can search it with nat
|
|
|
18
18
|
5. https://youtube.com/live/j7rItPwWDaY litdb and Jupyter Lab
|
|
19
19
|
6. https://youtube.com/live/SUtvtc7l6y0 litdb + GPT enhancements
|
|
20
20
|
7. https://youtube.com/live/3FZ1ROnCC6Y litdb + LiteLLM and streamlit
|
|
21
|
+
8. https://www.youtube.com/live/IKKTQSTXQmc litdb + Youtube and audio
|
|
22
|
+
9. https://youtube.com/live/MEf9rPI0Z1M litdb + Image search with text and image queries using CLIP
|
|
21
23
|
|
|
22
24
|
** installation
|
|
23
25
|
|
|
@@ -7,7 +7,7 @@ where = ["src"]
|
|
|
7
7
|
|
|
8
8
|
[project]
|
|
9
9
|
name = "litdb"
|
|
10
|
-
version = "2.1.
|
|
10
|
+
version = "2.1.4"
|
|
11
11
|
authors = [
|
|
12
12
|
{name = "John Kitchin", email = "jkitchin@andrew.cmu.edu"},
|
|
13
13
|
]
|
|
@@ -61,7 +61,12 @@ dependencies = [
|
|
|
61
61
|
"webvtt-py",
|
|
62
62
|
|
|
63
63
|
# the streamlit app
|
|
64
|
-
"streamlit"
|
|
64
|
+
"streamlit",
|
|
65
|
+
|
|
66
|
+
# image clip search
|
|
67
|
+
"pyperclip",
|
|
68
|
+
"PIL",
|
|
69
|
+
"pillow_heif"
|
|
65
70
|
]
|
|
66
71
|
|
|
67
72
|
[project.scripts]
|
|
@@ -344,14 +344,17 @@ The following subcommands can be used:
|
|
|
344
344
|
n = 3
|
|
345
345
|
|
|
346
346
|
# Now we do RAG on litdb if needed.
|
|
347
|
+
# I don't use the references below, so we do not assign them.
|
|
347
348
|
if rag:
|
|
348
|
-
rag_content,
|
|
349
|
+
rag_content, _ = get_rag_content(prompt, n)
|
|
349
350
|
else:
|
|
350
|
-
rag_content,
|
|
351
|
+
rag_content, _ = (None, None)
|
|
351
352
|
|
|
352
353
|
if rag and rag_content:
|
|
353
|
-
prompt = f"""
|
|
354
|
-
the user prompt. Reference the
|
|
354
|
+
prompt = f"""You are a professional scientist. Use the following
|
|
355
|
+
retrieved information to respond to the user prompt. Reference the
|
|
356
|
+
citations if you use that information. Each reference is surrounded by
|
|
357
|
+
<citation> tags. If the information is not relevent, do not use it.
|
|
355
358
|
|
|
356
359
|
<retrieved information>
|
|
357
360
|
{rag_content}
|
|
@@ -373,5 +376,5 @@ The following subcommands can be used:
|
|
|
373
376
|
|
|
374
377
|
# We don't always have data here, if you use your own rag data
|
|
375
378
|
|
|
376
|
-
if references:
|
|
377
|
-
|
|
379
|
+
# if references:
|
|
380
|
+
# richprint("References:\n\n" + references)
|
|
@@ -31,6 +31,7 @@ import tabulate
|
|
|
31
31
|
from tqdm import tqdm
|
|
32
32
|
import webbrowser
|
|
33
33
|
|
|
34
|
+
|
|
34
35
|
from .utils import get_config, init_litdb
|
|
35
36
|
from .db import get_db, add_source, add_work, add_author, update_filter, add_bibtex
|
|
36
37
|
from .openalex import get_data, get_text
|
|
@@ -38,6 +39,7 @@ from .pdf import add_pdf
|
|
|
38
39
|
from .bibtex import dump_bibtex
|
|
39
40
|
from .youtube import get_youtube_doc
|
|
40
41
|
from .audio import is_audio_url, get_audio_text, record
|
|
42
|
+
from .images import add_image, image_query, image_extensions
|
|
41
43
|
|
|
42
44
|
with warnings.catch_warnings():
|
|
43
45
|
warnings.filterwarnings("ignore", category=UserWarning, module="pydantic")
|
|
@@ -110,6 +112,7 @@ def add(
|
|
|
110
112
|
These are one time additions.
|
|
111
113
|
|
|
112
114
|
"""
|
|
115
|
+
|
|
113
116
|
for source in tqdm(sources):
|
|
114
117
|
# a work
|
|
115
118
|
if source.startswith("10.") or "doi.org" in source:
|
|
@@ -121,6 +124,9 @@ def add(
|
|
|
121
124
|
|
|
122
125
|
add_work(source, references, citing, related)
|
|
123
126
|
|
|
127
|
+
elif os.path.splitext(source)[1].lower() in image_extensions:
|
|
128
|
+
add_image(source)
|
|
129
|
+
|
|
124
130
|
# works from an author
|
|
125
131
|
elif "orcid" in source or source.lower().startswith("https://openalex.org/a"):
|
|
126
132
|
add_author(source)
|
|
@@ -550,7 +556,11 @@ def audio(playback=False):
|
|
|
550
556
|
@click.option("-e", "--emacs", is_flag=True, default=False)
|
|
551
557
|
@click.option("-i", "--iterative", is_flag=True, default=False)
|
|
552
558
|
@click.option("-m", "--max-steps", default=None)
|
|
553
|
-
@click.option(
|
|
559
|
+
@click.option(
|
|
560
|
+
"-f",
|
|
561
|
+
"--fmt",
|
|
562
|
+
default=(" {{ i }}. ({{ score|round(3) }})" " {{ source }}\n{{ text[:200] }}\n\n"),
|
|
563
|
+
)
|
|
554
564
|
@click.option("-x", "--cross-encode", is_flag=True, default=False)
|
|
555
565
|
def vsearch(query, n, emacs, fmt, cross_encode, iterative, max_steps):
|
|
556
566
|
"""Do a vector search on QUERY.
|
|
@@ -697,6 +707,14 @@ def chat_command(model, debug):
|
|
|
697
707
|
cli.add_command(chat_command, name="chat")
|
|
698
708
|
|
|
699
709
|
|
|
710
|
+
@cli.command()
|
|
711
|
+
@click.argument("query", nargs=-1)
|
|
712
|
+
@click.option("-c", "--clipboard", is_flag=True, default=False)
|
|
713
|
+
@click.option("-n", default=1, help="Number of hits to retrieve")
|
|
714
|
+
def image_search(query, clipboard, n):
|
|
715
|
+
image_query(" ".join(query), clipboard, n)
|
|
716
|
+
|
|
717
|
+
|
|
700
718
|
@cli.command()
|
|
701
719
|
@click.argument("source")
|
|
702
720
|
@click.option("-f", "--fmt", default=None)
|
|
@@ -1421,7 +1439,7 @@ def suggest_reviewers(query, n):
|
|
|
1421
1439
|
lki = d.get("last_known_institutions", [])
|
|
1422
1440
|
if lki == []:
|
|
1423
1441
|
affils = d.get("affiliations", [])
|
|
1424
|
-
if len(affils) >=
|
|
1442
|
+
if len(affils) >= 1:
|
|
1425
1443
|
lki = affils[0]["institution"]["display_name"]
|
|
1426
1444
|
else:
|
|
1427
1445
|
lki = "unknown"
|
|
@@ -51,7 +51,8 @@ def get_db():
|
|
|
51
51
|
)
|
|
52
52
|
|
|
53
53
|
db.execute(
|
|
54
|
-
"""create virtual table if not exists
|
|
54
|
+
"""create virtual table if not exists
|
|
55
|
+
fulltext using fts5(source, text)"""
|
|
55
56
|
)
|
|
56
57
|
|
|
57
58
|
db.execute(
|
|
@@ -94,6 +95,24 @@ def get_db():
|
|
|
94
95
|
prompt_history(rowid integer primary key,
|
|
95
96
|
prompt text)"""
|
|
96
97
|
)
|
|
98
|
+
|
|
99
|
+
# For images
|
|
100
|
+
model = SentenceTransformer("clip-ViT-B-32")
|
|
101
|
+
_, dim = model.encode(["test"]).shape
|
|
102
|
+
|
|
103
|
+
db.execute(
|
|
104
|
+
f"""create table if not exists
|
|
105
|
+
images(rowid integer primary key,
|
|
106
|
+
source text unique,
|
|
107
|
+
embedding F32_BLOB({dim}),
|
|
108
|
+
date_added text)"""
|
|
109
|
+
)
|
|
110
|
+
|
|
111
|
+
db.execute(
|
|
112
|
+
"""create index if not exists image_idx
|
|
113
|
+
ON images (libsql_vector_idx(embedding))"""
|
|
114
|
+
)
|
|
115
|
+
|
|
97
116
|
return db
|
|
98
117
|
|
|
99
118
|
|
|
@@ -0,0 +1,73 @@
|
|
|
1
|
+
"""Add-on to index and search images by text.
|
|
2
|
+
|
|
3
|
+
This uses the clip transformer to get embedding vectors of images. Then you can
|
|
4
|
+
search the images either by textual descriptions, or by using similar images.
|
|
5
|
+
|
|
6
|
+
"""
|
|
7
|
+
|
|
8
|
+
from .db import get_db
|
|
9
|
+
from sentence_transformers import SentenceTransformer
|
|
10
|
+
from PIL import Image, ImageGrab
|
|
11
|
+
from pillow_heif import register_heif_opener
|
|
12
|
+
import numpy as np
|
|
13
|
+
import datetime
|
|
14
|
+
import os
|
|
15
|
+
import pyperclip
|
|
16
|
+
|
|
17
|
+
register_heif_opener()
|
|
18
|
+
|
|
19
|
+
image_extensions = Image.registered_extensions().keys()
|
|
20
|
+
|
|
21
|
+
model = SentenceTransformer("clip-ViT-B-32")
|
|
22
|
+
db = get_db()
|
|
23
|
+
|
|
24
|
+
|
|
25
|
+
def add_image(path):
|
|
26
|
+
"""Embed and add the image in path to db."""
|
|
27
|
+
|
|
28
|
+
emb = model.encode(Image.open(path))
|
|
29
|
+
|
|
30
|
+
q = """insert or ignore into
|
|
31
|
+
images(source, embedding, date_added)
|
|
32
|
+
values (?, ?, ?)"""
|
|
33
|
+
db.execute(
|
|
34
|
+
q,
|
|
35
|
+
(
|
|
36
|
+
os.path.abspath(path),
|
|
37
|
+
emb.astype(np.float32).tobytes(),
|
|
38
|
+
datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
|
|
39
|
+
),
|
|
40
|
+
)
|
|
41
|
+
db.commit()
|
|
42
|
+
|
|
43
|
+
|
|
44
|
+
def image_query(query=None, clipboard=False, n=1):
|
|
45
|
+
"""Query the image database.
|
|
46
|
+
|
|
47
|
+
If query is not None, it is either a text description or path to image.
|
|
48
|
+
If clipboard is True, use an image or text from the clipboard
|
|
49
|
+
"""
|
|
50
|
+
if query and os.path.exists(query):
|
|
51
|
+
emb = model.encode(Image.open(query))
|
|
52
|
+
elif query and not os.path.exists(query):
|
|
53
|
+
emb = model.encode(query)
|
|
54
|
+
elif clipboard:
|
|
55
|
+
clip = ImageGrab.grabclipboard()
|
|
56
|
+
if isinstance(clip, Image.Image):
|
|
57
|
+
emb = model.encode(clip)
|
|
58
|
+
else:
|
|
59
|
+
clip = pyperclip.paste()
|
|
60
|
+
print(clip)
|
|
61
|
+
emb = model.encode(clip)
|
|
62
|
+
|
|
63
|
+
emb = emb.astype(np.float32).tobytes()
|
|
64
|
+
|
|
65
|
+
q = """select source from vector_top_k('image_idx', ?, ?)
|
|
66
|
+
join images on images.rowid = id"""
|
|
67
|
+
results = db.execute(q, (emb, n)).fetchall()
|
|
68
|
+
|
|
69
|
+
for (row,) in results:
|
|
70
|
+
print(row)
|
|
71
|
+
Image.open(row).show()
|
|
72
|
+
|
|
73
|
+
return results
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
Metadata-Version: 2.2
|
|
2
2
|
Name: litdb
|
|
3
|
-
Version: 2.1.
|
|
3
|
+
Version: 2.1.4
|
|
4
4
|
Summary: A literature database tool with GPT integration.
|
|
5
5
|
Author-email: John Kitchin <jkitchin@andrew.cmu.edu>
|
|
6
6
|
Maintainer-email: John Kitchin <jkitchin@andrew.cmu.edu>
|
|
@@ -40,6 +40,9 @@ Requires-Dist: IPython
|
|
|
40
40
|
Requires-Dist: yt_dlp
|
|
41
41
|
Requires-Dist: webvtt-py
|
|
42
42
|
Requires-Dist: streamlit
|
|
43
|
+
Requires-Dist: pyperclip
|
|
44
|
+
Requires-Dist: PIL
|
|
45
|
+
Requires-Dist: pillow_heif
|
|
43
46
|
|
|
44
47
|
#+title: litdb - a literature and document database
|
|
45
48
|
|
|
@@ -61,6 +64,8 @@ The idea is you add papers to your database, and then you can search it with nat
|
|
|
61
64
|
5. https://youtube.com/live/j7rItPwWDaY litdb and Jupyter Lab
|
|
62
65
|
6. https://youtube.com/live/SUtvtc7l6y0 litdb + GPT enhancements
|
|
63
66
|
7. https://youtube.com/live/3FZ1ROnCC6Y litdb + LiteLLM and streamlit
|
|
67
|
+
8. https://www.youtube.com/live/IKKTQSTXQmc litdb + Youtube and audio
|
|
68
|
+
9. https://youtube.com/live/MEf9rPI0Z1M litdb + Image search with text and image queries using CLIP
|
|
64
69
|
|
|
65
70
|
** installation
|
|
66
71
|
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|