Incorrect format validation, number or array?

[whatwoot]

I found that the prompt specifies the format of number：

Your output should strictly conforms the following json format without any additional contents: {"summary_reason": string, "short_memory_index": number, "middle_memory_index": number, "long_memory_index": number, "reflection_memory_index": number}

But array is needed in guardrails?

---

I was given the following JSON response, which had problems due to incorrect values.

{
"summary_reason": "The decision to invest in TSLA based on the observed price difference from the current to the next trading day is influenced by short-term factors.",
"short_memory_index": -1,
"middle_memory_index": -1,
"long_memory_index": -1,
"reflection_memory_index": -1
}

Help me correct the incorrect values based on the given error messages.

Error Messages:
"JSON does not match schema:\n{\n "$.reflection_memory_index": [\n "-1 is not of type 'array'"\n ],\n "$.long_memory_index": [\n "-1 is not of type 'array'"\n ],\n "$.middle_memory_index": [\n "-1 is not of type 'array'"\n ],\n "$.short_memory_index": [\n "-1 is not of type 'array'"\n ]\n}"

[morning-star-001]

The generated data for the Fake-Sample-Data is not clear, Do you mind elaborating on how you were able to generate data for use?

A guide for instance on generating tsla.pkl

[Carolyn-Jiang]

The output files generated by the script “05-get_sentiment_by_ticker.py” are named with the prefix “subset_symbols_” followed by the corresponding ticker name. These files serve as the final input for FinMem. To use them, you can simply rename the file to “<ticker name>.pkl”. You can use this model via Gitpod, please refer to this video. https://github.com/pipiku915/FinMem-LLM-StockTrading/blob/main/FinMem.mp4 Get Outlook for iOS<https://aka.ms/o0ukef>

…

________________________________ From: morning-star-001 ***@***.***> Sent: Friday, December 27, 2024 12:18:53 PM To: pipiku915/FinMem-LLM-StockTrading ***@***.***> Cc: Subscribed ***@***.***> Subject: Re: [pipiku915/FinMem-LLM-StockTrading] Incorrect format validation, number or array? (Issue #25) The generated data for the Fake-Sample-Data is not clear, Do you mind elaborating on how you were able to generate data for use? A guide for instance on generating tsla.pkl — Reply to this email directly, view it on GitHub<#25 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ANUUDRW3FFWERWPYBTFF4UD2HTIK3AVCNFSM6AAAAABUEPDVIWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNRTGI4TONJQG4>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[morning-star-001]

thanks for the response..

Got this error

FileNotFoundError: [Errno 2] No such file or directory: './data/05_env_data/env_data.pkl'

Although I created the folder and used the env_data.pkl found in the Fake-Sample-Data, and got another error response:

Ticker list: ------ ['BAC', 'DIS', 'GM', 'MRNA', 'NVDA', 'PFE'] 0%| | 0/496 [00:00<?, ?it/s] Traceback (most recent call last): File "/llm/llm/./data-pipeline/05-get_sentiment_by_ticker.py", line 150, in <module> export_sub_symbol(cur_symbol_lst, senti_model_type = 'FinBERT') File "/llm/llm/./data-pipeline/05-get_sentiment_by_ticker.py", line 132, in export_sub_symbol new_dict, ticker_dict_byDate = subset_symbol_dict(input_dir, cur_symbol_0) File "/llm/llm/./data-pipeline/05-get_sentiment_by_ticker.py", line 28, in subset_symbol_dict cur_filing_q = v[2]['filling_q'] # form q KeyError: 'filling_q'

[morning-star-001]

The above error has been debugged, been a typo in the env_data.pkl itself In the filling_q key.

A new error did surface though.

Traceback (most recent call last): File "/llm/llm/./data-pipeline/05-get_sentiment_by_ticker.py", line 151, in <module> export_sub_symbol(cur_symbol_lst, senti_model_type = 'FinBERT') File "/llm/llm/./data-pipeline/05-get_sentiment_by_ticker.py", line 137, in export_sub_symbol print('finBERT Date" 2023-05-30: ----- ', new_dict[datetime.date(2023, 5, 30)]['news']) KeyError: datetime.date(2023, 5, 30)

[Carolyn-Jiang]

This is because fake data does not contain data for this day. You can comment out all prints.

…

________________________________ From: morning-star-001 ***@***.***> Sent: Friday, December 27, 2024 11:15:56 PM To: pipiku915/FinMem-LLM-StockTrading ***@***.***> Cc: Yuechen Jiang ***@***.***>; Comment ***@***.***> Subject: Re: [pipiku915/FinMem-LLM-StockTrading] Incorrect format validation, number or array? (Issue #25) The above error has been debugged, been a typo in the env_data.pkl itself In the filling_q key. A new error did surface though. Traceback (most recent call last): File "/llm/llm/./data-pipeline/05-get_sentiment_by_ticker.py", line 151, in <module> export_sub_symbol(cur_symbol_lst, senti_model_type = 'FinBERT') File "/llm/llm/./data-pipeline/05-get_sentiment_by_ticker.py", line 137, in export_sub_symbol print('finBERT Date" 2023-05-30: ----- ', new_dict[datetime.date(2023, 5, 30)]['news']) KeyError: datetime.date(2023, 5, 30) — Reply to this email directly, view it on GitHub<#25 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ANUUDRRAG42YI2BYLO3NCBD2HVVKZAVCNFSM6AAAAABUEPDVIWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNRTG44TCOJSG4>. You are receiving this because you commented.Message ID: ***@***.***>

[Carolyn-Jiang]

Please refer to https://coling2025cryptotrading.thefin.ai And https://github.com/felis33/coling-cryptocurrency-trading-challenge You can find open-source data and start kit here.

…

________________________________ From: morning-star-001 ***@***.***> Sent: Saturday, December 28, 2024 1:22:15 AM To: pipiku915/FinMem-LLM-StockTrading ***@***.***> Cc: Yuechen Jiang ***@***.***>; Mention ***@***.***> Subject: Re: [pipiku915/FinMem-LLM-StockTrading] Incorrect format validation, number or array? (Issue #25) @Carolyn-Jiang<https://github.com/Carolyn-Jiang> Could you helo with the price_data.parquet file please — Reply to this email directly, view it on GitHub<#25 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ANUUDRSZK4SZYOI5RLJAKVD2HWEEPAVCNFSM6AAAAABUEPDVIWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNRTHA4TAMZYHE>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[whatwoot]

@Carolyn-Jiang

Any idea for my question？Thank u a lot！

I found that the prompt specifies the format of number：

Your output should strictly conforms the following json format without any additional contents: {"summary_reason": string, "short_memory_index": number, "middle_memory_index": number, "long_memory_index": number, "reflection_memory_index": number}

But array is needed in guardrails?

I was given the following JSON response, which had problems due to incorrect values.

{ "summary_reason": "The decision to invest in TSLA based on the observed price difference from the current to the next trading day is influenced by short-term factors.", "short_memory_index": -1, "middle_memory_index": -1, "long_memory_index": -1, "reflection_memory_index": -1 }

Help me correct the incorrect values based on the given error messages.

Error Messages: "JSON does not match schema:\n{\n "$.reflection_memory_index": [\n "-1 is not of type 'array'"\n ],\n "$.long_memory_index": [\n "-1 is not of type 'array'"\n ],\n "$.middle_memory_index": [\n "-1 is not of type 'array'"\n ],\n "$.short_memory_index": [\n "-1 is not of type 'array'"\n ]\n}"