Update machine report classes

[anacso17]

• Add manual failure annotation for October 18th problem
  • We had problems with PPS in this period (user shift without beam and with wrong machine shift annotations).
• Do not ignore wrong shift annotations in the beggining of an analized period
  • There is a feature that was introduced to ignore wrong shift annotations within 1min (for example, to not indicate user shift problems if the MachShift mode PV changes in value 1min later than scheduled). This commit changes this feature to not ignore wrong shift annotations in the beggining of an analized period. This was causing a wrong indication of transition from "correct shift" to "wrong shift" when the analized period started with a wrong shift and, as a consequence, causing a wrong beam dump annotation.
• Update example of code to generate a report

[xresende]

@anacso17 , did you notice that integration tests are failing? I looked at the output bu could not understand what is going on. maybe it is a temporary service instability?

[anacso17]

@anacso17 , did you notice that integration tests are failing? I looked at the output bu could not understand what is going on. maybe it is a temporary service instability?

Hi @xresende , this is because of the fail in mathphys wheel build, which will be fixed with this PR lnls-fac/mathphys#37.