You make a number of major errors in your premises.
You use the phrase "spontaneous generation" in a manner where it can be taken two ways. One is the spontaneous generation of a complex multicellular organism from material which is chemically and energetically inert, as far as producing complex molecules is concerned. The second way is the interpretation used by science when considering Abiogenesis and that is the gradual build up of complexity from a basic molecule, which itself is the product of natural chemical bonding and sufficient available energy, through imperfect self replication.
You are using the phrase knowing that it has been correctly used by scientists to define undirected origins but are aware that most anti-evos will take it as used in the first definition. Simply put, the spontaneous generation of science is not the same spontaneous generation Pasteur was concerned about and disproved. Your attempt to conflate the two is simple misdirection.
We know that the basic building blocks of DNA and RNA can be produced through the application of energy in the right conditions, both through lab experiments and the discovery of clouds of amino acids, sugars and alcohols in space. However it may be that the type of molecules which gave rise to Earthly life were not originally DNA or RNA, those may have come much later. We simply do not know.
This current lack of knowledge does not indicate whether we will or will not know the answers and it certainly does not excuse your attempt to poison the well through equivocal use of spontaneous generation.
Your second point that Evolution expects one species to miraculously change into another in the lab is another bit of subterfuge. By using the word 'miraculous' you are attempting to link common descent to religion without providing evidence that your premise is correct. You are in effect assuming your conclusion.
As for one species giving birth to another that is neither expected from, nor required by, Evolution. It is a strawman derived from a floating definition of 'species' common among anti-evolutionists.
A species, which is a human construct designed to give us classifications where nature is at times muddled, is defined by Science a few different way, the most common is the cessation of gene flow from between two groups that were at one time a single population. This type of speciation has been observed in the wild, primarily among plants but has also been observed in animals and insects. We have also observed situations where a population is in the process of speciating such as the Asian Greenish Warbler and other ring species. We see the fine speciation in extant species that anti-evolutionists are demanding to see in the fossil record.
We have also witnessed speciation in the lab, including with "Drosophila melanogaster which are in the Order Diptera or an insect with two wings. Through what some have labeled 'macro-mutation' a new species of fly was produced having four wings so should rightly be re-classed to belong to one of the multi-winged Orders, perhaps Neuroptera. This extreme version of speciation in the lab is far from the only example.
The concept of speciation usually adopted by anti-evolutionists is quite different from that of science. To these antievolutionists a speciation event would be one where a species changed radically and abruptly to a species with a dramatically different morphology.
It is easy to understand why this 'saltational' event is common among anti-volutionists, it can at times be very difficult to see the difference between two closely related species but very easy to see the the difference between two more distantly related species. However, that distance, while easy to see, misses a great many steps that had to be taken by *each* species to get where they are from their common ancestor.
However that definition of *speciation* is not used by scientist and is not required or expected by the SToE. It is a strawman version of what the SToE does predict, a gradual, stepwise divergence of two species from their common ancestor and from each other over time. This is not to claim that evolution proceeds at a constant pace through time, there is no set rate of change. Evolution proceeds much like a faltering drunk, a few steps forward, a step back, a few steps sideways but inevitably a new species emerges just as the drunk will eventually step off the sidewalk.
At no time is it expected or required for a member of a species to give birth to different species. There are *no* cats giving birth to dogs, or fish giving birth to amphibians, Evolution has never claimed that occurs. The only people who claim that are anti-evolutionists.
The third bad premise you use is the requirement that only direct observations are acceptable by science. This is far from the truth, science not only uses indirect observation but uses it more than direct observation.
Aside from the common use of indirect observation and inferential conclusions by all sciences, many mechanisms can be broken down into smaller more manageable bites testable in the lab. For Evolution, we can test the efficacy of mutations to provide allele variation, we can test the inheritability of alleles, we can test the inheritability of mutations, we can test for the probabilities for beneficial, neutral and deleterious mutations, we can test for the ability of a number of different types of selection to 'fix' an allele in a population, we can compare two or more genomes to find, not just similarities but differences in both the functional DNA and the junk DNA to determine common heritage. There are many tenets of the SToE which can and have been tested in the lab and observed in the wild, we do not need to observe more than the tracks left in the sand by past events to determine how they came to be.
This is no different than any other branch of science.