pashpashpash / vault-ai

OP Vault ChatGPT: Give ChatGPT long-term memory using the OP Stack (OpenAI + Pinecone Vector Database). Upload your own custom knowledge base files (PDF, txt, epub, etc) using a simple React frontend.
https://vault.pash.city
MIT License
3.25k stars 306 forks source link

No Summary of content, just of style #95

Open CWWhitney opened 4 weeks ago

CWWhitney commented 4 weeks ago

The description the process sounds perfect for my application so I am testing out the paid version. I loaded roughly 1000 page .txt file to Summarize.wtf. It took about a minute but did not return a summary of the content of the .txt.

About the session (all I could find): Your Usage this Month 1/200 Questions 3.15 MB/1600 MB Uploaded 1101/10000 Pages Plus Tier Subscription

TOTAL TOKENS USED: 13031

Screenshot 2024-06-17 at 1 59 42 PM

The return from the interface

The input text consists of various chunks, each containing information about different aspects of a document or article. Here is a breakdown of the key points from each chunk:

Chunk ID: 5181699561357312-0

The chunk contains formatting information for the document.
Chunk ID: 5181699561357312-1

Provides information about the item type and note associated with it.
Chunk ID: 5181699561357312-2

Indicates the date added and modified for the item.
Chunk ID: 5181699561357312-3

Provides additional information about the modification date.
Chunk ID: 5181699561357312-4

Contains annotations and quotes related to public health policy.
Chunk ID: 5181699561357312-12

Discusses the need for more research on interventions related to public health policy in schools.
Chunk ID: 5181699561357312-56

Talks about the concept of placekeeping and its role in creating equitable local food systems.
Chunk ID: 5181699561357312-57

Provides information about the item type, which is a blog post.
Chunk ID: 5181699561357312-58

Gives the author's name for the blog post.
Chunk ID: 5181699561357312-60

Provides information about the language in which the blog post is written.
Chunk ID: 5181699561357312-63

Indicates the date added and modified for the blog post.
Chunk ID: 5181699561357312-64

Provides additional information about the modification date.
Chunk ID: 5181699561357312-72

Discusses the school food environment and its impact on children's dietary behaviors.
Chunk ID: 5181699561357312-74

Presents an abstract for a journal article related to community food security through gardening.
Chunk ID: 5181699561357312-75

Talks about the author's name for the journal article.
Chunk ID: 5181699561357312-78

Provides a brief overview of a conference, including the number of attendees and papers presented.
Chunk ID: 5181699561357312-110

Indicates the item type as a journal article.
Chunk ID: 5181699561357312-113

Provides the author's name for the journal article.
Chunk ID: 5181699561357312-132

Gives a summary of the abstract related to nutrition interventions in the Philippines.
Chunk ID: 5181699561357312-147

Provides information about the number of abstracts submitted and presentations at a conference.
Chunk ID: 5181699561357312-160

Lists the tags associated with the item.
Chunk ID: 5181699561357312-175

Gives the authors' names for a journal article on addressing community food security through gardening.
Chunk ID: 5181699561357312-110

Provides formatting information for the document.
These chunks provide details about the type of item, authorship, dates, abstracts, and tags associated with the items. Overall, the input text provides a comprehensive understanding of the different aspects covered in the document or article
CWWhitney commented 4 weeks ago

I tried a second time with an unformatted .txt file and got a more detailed return. You may need a check on the formatting of the document or a reformatting step so that the 'chunks' are not the focus of the process.