OpenEuphoria

1. Thankful to petelomax and Phix

• Posted by jesito Feb 16, 2023
• 824 views

I´ve been involved in Euphoria since the beginnings. I discovered long ago RDS Euphoria and found it easy to learn and very productive for my way of working.

I developed several applications, among them a website generator for a friend. He is a documentalist. He has devoted more than 25 years to look for documents related to the Monaco de Bacardí family in the northeastern Spain. He uses to go to villages, small and big cities visiting parishes, notaries and civil registries. He takes pictures of the relevant documents, then back at home, manually types those documents in big MSWord documents. Most of these documents are handwrited, in latin, old catalanish and old spanish. No way of using OCR. So far he has been able to recover around 10700 documents.

The work starts converting these big documents (10 in total currently) from MSWord to UTF-8 text. Then a parser written in Euphoria analyzes these documents, checking some marking rules that we agreed to diferentiate documents, notaries names and actors names. The output of this phase is given back to him until everything is fine.

The second phase is another program that reads the big text files and does a preprocess to split them in separate text files (to be able to be amended if necessary) adding some primary HTML info and creating a SQL database with information to be able to cross reference the documents.

The last program builds the website adding CSS info and creating the document index and the web pages.

For years we have been doing one or two builds per year. But the pandemics stopped this (he wasn't allowed to travel as it used to). I have to say that my friend is a real nerd in computers, so when I gave him a flashdrive with the programs and a backup of the site, he didn't reaalize the importance of having it stored in a safe place. The last update was in 2018, before the pandemic. I also kept a backup copy on my harddrive, with the Euphoria environment.

Recently he called me asking for an update. He had included many new documents and wanted to update the site. I asked him about the flashdrive (my harddisk broke down some time ago) and he didn't remember about it... Hopefully I got a previous version of 2014 that was a starting point. There were many difficulties to start again. I moved long ago to Linux and the apps were written and run under WinXP/Win7.

I tried at first to run the old apps under linux, but there were some dependencies (sqlite3, etc) not easy to meet. To ease the work I set up a Win10 PC and loaded the old programs. They run smoothly with minor flaws. But the generated web was a very old one with archaic design. We evolved it a lot from 2014 to 2018. So I took the 2018 website (hosted in a 3rd party provider) and looked how the pages were built. Following that I was able to recover the structure and physical aspect of the site.

I didn't remember either how I wrote the programs, it was so long ago. Also I forgot a bit about Euphoria. (I'm an old chap, currently retired and my mind is not as good as it used to be).So I had to start from the beginning again. After some time I was able to solve the first phase and gave him the report on the errors in his documents, that he promptly checked and corrected.

For the second phase I found the process painfully slow. It used to take 24 hours to process the 10.700 documents. Looking for a quick solution I recalled having read about Phix, that was stated as much faster than Euphoria and highly compatible. So I tried to install Phix on linux (I'm allergic to Windows after a professional life with plenty of it). I got some problems at first because the segmentation fault problem in Linux, but there is a workaraound that works and started to work with it. Same problems than before, so I decided to move again to MSWindows and finish the work there.

Phix installed like a breeze, and when I run the second phase, it took only 1 hour and a half where Eu was taking from 22 to 24 hours. A big surprise and a big help, I was able to run the entire process many times per day, shortening the development time a lot.

I got a few compatibility problems, most of them were routines that Phix already included and I had only to comment on the old sources. The work it's still not finished but we're very close to. The 10.700 documents are propely generated, now only remain some bells and whistles from the document index and the entry page. I'll post a link to the page once finished for the case someone is interested.

I wanted to say a big thanks to Pete Lomax for Phix, and also a big thanks to Robert Craig for the language and to the Euphoria community for the unpayable and outstanding support and help given all those years.

Jes.