Loop summary generation

[lebonq]

I've added miniflux-ai to my docker compose of miniflux. Everything work fine except the news summary, every 2–3 minutes the fetch_unread_entries.py is called and generate a summary for every unread entry. However, it runs multiple times on the same article as this :

Mini-flux-ai should be able to know if an article as already being summarized to prevent useless use of the model and avoid to break the layout of the article.

[Qetesh]

Thanks for using and feedback

The entry_entry method of entry_filter.py is used in the project to filter summarized articles by title and style_block style keywords in the config file.

Could you provide the configuration file for title in agents, and logs to identify the issue

[lebonq]

I extracted the entries.json to see what is wrong.

[
    {
        "datetime": "2025-02-18T09:54:13.595325+01:00",
        "category": "Actualités",
        "title": "Guerre en Ukraine : \"Les Européens sont en train de négocier avec les Américains\", assure Nicole Gnesotto, vice-présidente de l'Institut Jacques-Delors",
        "content": "Nicole Gnesotto assure que les Européens négocient avec les Américains. \nIl y a deux négociations : américano-russe et euro-américaine.\nL'Europe doit être associée aux négociations sur l'Ukraine."
    },
    {
        "datetime": "2025-02-18T09:54:13.595325+01:00",
        "category": "Actualités",
        "title": "Guerre en Ukraine : \"Les Européens sont en train de négocier avec les Américains\", assure Nicole Gnesotto, vice-présidente de l'Institut Jacques-Delors",
        "content": "Les Européens négocient avec les Américains sur l'Ukraine.\nL'Europe doit être associée aux négociations pour un accord de paix.\nIl y a deux négociations : américano-russe et euro-américaine."
    }
]

And the title in agent is defined as follow :

 title: '<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 17.777 14.283" width="17.777" height="14.283"> <style> path { fill: #4b4b4b; } @media (prefers-color-scheme: dark) { path { fill: #d0d0d0; } } </style> <g fill="#d0d0d0" fill-opacity="1" transform="translate(2.261,-1.754)"> <path d="M-2.261 3.194v6.404c0 1.549 0.957 4.009 4.328 4.188h9.224l0.061 1.315c0.04 0.882 0.663 1.222 1.205 0.666l2.694-2.356c0.353-0.349 0.353-0.971 0-1.331L12.518 10.047c-0.525-0.524-1.205-0.196-1.205 0.665v1.091H2.257c-0.198 0-2.546 0.221-2.546-2.911V3.194c0-0.884-0.362-1.44-0.99-1.44-1.106 0-0.956 1.439-0.982 1.44z"/> </g> <path d="M5.679 1.533h8.826c0.421 0 0.753-0.399 0.755-0.755 0.002-0.36-0.373-0.774-0.755-0.774H5.679c-0.536 0-0.781 0.4-0.781 0.764 0 0.418 0.289 0.764 0.781 0.764zm0 4.693h4.502c0.421 0 0.682-0.226 0.717-0.742 0.03-0.44-0.335-0.787-0.717-0.787H5.679c-0.402 0-0.763 0.214-0.781 0.71-0.019 0.535 0.379 0.818 0.781 0.818z" fill="#d0d0d0"/> </svg> AI summary:'

I checked the code and I think process_entry should store also the id field of entry, thus entry_filter could filter using the unique ID of each entry instead of trying to match a string that represent the title.

[Qetesh]

If a unique ID needs to be saved, data persistence and the database must be considered, and since this project will perform a hotfix on miniflux and there will be two databases, I will consider this solution based on the future.

Based on the configuration file, there shouldn't be any issues

Considering that fetch_unread_entries.py is normally executed once a minute without webhook callback configuration, this does not match the interval of every 2–3 minutes you mentioned

Need to check application stdout logs to check this issue

[lebonq]

I check closely the log, it does happen every minute. Sorry about the imprecision